Regulating Facial Recognition Technology to Reduce Bias

By Khortlan Becton, JD, MTS, The Restorative Education Institute (1)

This blog post is an excerpt from the 2023 Hooks Institute Policy Papers “The Promise and Peril: Unpacking the Impact of A.I. and Automation on Marginalized Communities.” Read more here.

I. Introduction

From unlocking a phone to identifying shoplifters in real-time, facial recognition technology (“FRT”) use is increasing among private companies and having an increasingly large impact on the public. According to one study, the global FRT market is expected to grow from $3.8 billion in 2020 to $8.5 billion by 2025 (MarketsandMarkets, 2020). Domestically, FRTs are a central aspect of artificial intelligence (“AI”) use and development in the U.S. pri- vate sector. The following statistics demonstrate the growing prevalence of FRT use in the U.S. private sector: 72% of hotel operators are expected to deploy FRTs by 2025 to identify and interact with guests; by 2023, 97% of airports will roll out FRTs; excluding Southwest Airlines, most major US airlines currently use FRTs (Calvello, 2019).

This explosion of private FRT use has prompted many professional organizations and community organizers to call for a moratorium on FRT use until the enactment of state and federal regulatory actions. One such group noted that industry and government have adopted FRTs “ahead of the development of principles and regulations to reliably assure their consistently appropriate and non-prejudicial use” (Association for Computing Machinery [ACM], 2020). Among the stakeholders calling for such moratoriums is a concern over the alarming level of bias present within commercial FRT systems. Given the widespread integration of FRTs throughout society, both presently and to come, the presence of bias in FRTs is particularly troublesome as decision-making driven by biased FRT can lead to significant physical and legal injuries. For example, self-driving cars are more likely to hit dark-skinned pedestrians (Samuel, 2019). Biased FRTs also have the likelihood of producing discriminatory hiring decisions, credit approvals, or mortgage approvals.

Though the observable and conceivable consequences of bias in FRTs are virtually boundless, state and federal regulatory schemes have not adapted to the growth of FRTs. A continuing lag in regulations designed to address bias in FRTs will likely lead to a range of discriminatory effects that existing agencies do not have the capacity to prevent or redress. Therefore, a federal regulatory scheme propagated by a new agency specifically authorized to regulate AI technologies will better ensure the governance of private entities’ use of facial recognition technologies to address bias than the current regulatory scheme.

A. The Relationship Between AI and FRTs

In popular usage, AI refers to the ability of a computer or machine to mimic the capabilities of the human mind and combining these and other capabilities to perform functions a human might perform (IBM, 2020). AI-powered machines are usually classified into two groups—general and narrow (Towards Data Science, 2018). Narrow AI, which drives most of the AI that surrounds us today, is trained and focused to perform specific tasks. (IBM, 2020). General AI is AI that more fully replicates the autonomy of the human brain—AI that can solve many types of problems and even choose the problems it wants to solve without human intervention (IBM, 2020).

Machine learning is a subset of AI application that enables an application to progressively reprogram itself, digesting data input by human users, to perform the specific task the application is designed to perform with increasingly greater accuracy (IBM, 2020). Deep learning, a subset of machine learning, allows applications to automatically identify the features to be used for classification, without human intervention (IBM, 2020).

Facial recognition technologies are artificial intelligence systems programmed to identify or verify the identity of a person using their face (Thales Group, 2021). “A general statement of the problem of machine recognition of faces can be formulated as follows: given still or video images of a scene, identify or verify one or more persons in the scene using a stored database of faces” (Chellappa et al., 2003). Face recognition is often described as a process that first involves four steps: face detection, face alignment, feature extraction, and face recognition (Brownlee, 2019).

  1. Face Detection. Locate one or more faces in the image with a bounding box.
  2. Face Alignment. Normalize the face to be consistent with the database, such as geometry and
  3. photometrics.
  4. Feature Extraction. Extract features from the face that can be used for the recognition task.
  5. Face Recognition. Perform matching of the face against one or more known faces in a prepared database (Brownlee, 2019).

Companies are developing and implementing FRTs in new and potentially beneficial ways, such as: helping news organizations identify celebrities in their coverage of significant events, providing secondary authentication for mobile applications, automatically indexing image and video files for media and entertainment companies, and allowing humanitarian groups to identify and rescue human trafficking victims (Amazon Web Services [AWS], 2021). Recently, FRT has been in the news for its application in the investigation of the Jan. 6, 2021, Capital riot (Sakin, 2021). Other news stories about facial recognition have centered on the coronavirus pandemic. One business proposed creating immunity passports for those who are no longer at risk of contracting or spreading COVID-19 and to use FRTs to identify the immunity passport holder (Sakin, 2021). A MarketsandMarkets (2020) study estimates that the global facial recognition market is expected to grow from $3.8 billion in 2020 to $8.5 billion by 2025.

The Federal Trade Commission’s (“FTC”) recent settlement with Everalbum, a California-based developer of a photo storage app, exemplifies the growth of FRT use in the commercial sector and the liabilities companies may face for implementing the technology. In its complaint, the FTC alleged that Everalbum, which offered an app that allowed users to upload photos and videos to be stored and organized, launched a new feature that, by default, used face recognition to group users’ photos by faces of the people who appear in the photos (Everalbum, Inc., n.d.). Everalbum also allegedly used, without affirmative express consent, users’ uploaded photos to train and develop its own FRT (Everalbum, Inc., n.d.). Regarding its implementation of FRTs, the FTC charged Everalbum
for engaging in unfair or deceptive acts or practices, in violation of Section 5(a) of the Federal Trade Commission Act, by misrepresenting that it was not using facial recognition unless the user enabled it or turned it on (Everal- bum, Inc., n.d.). In January 2021, Everalbum settled the FTC allegations concerning its deceptive use of FRTs. The proposed settlement requires Everalbum to delete all face embeddings the company derived from photos of users who did not give their express consent to their use and any facial recognition models or algorithms developed with users’ photos or videos (Everalbum, Inc., n.d.). The company must also obtain a user’s express consent before using biometric information it collected from the user to create face embeddings or develop FRTs (Everalbum, Inc., n.d.). Everalbum’s recent settlement with the FTC underscores the nascency of federal governance of FRTs, as the Everalbum settlement is among the first of few federal agency enforcements targeting commercial use of FRTs (Federal Trade Commission [FTC], 2019) (2). Signaling the potential for increasing regulation and enforcement in this area, FTC Commissioner Rohit Chopra noted that FRT “is fundamentally flawed and reinforces harmful biases” while highlighting the importance of “efforts to enact moratoria or otherwise severely restrict its use” (Federal Trade Commission [FTC], (2021).

B. Bias in Facial Recognition Technologies

Although proponents of FRTs boast high accuracy rates, a growing body of research exposes divergent error rates in FRT use across demographic groups (Najibi, 2020). In the landmark 2018 “Gender Shades” report, MIT and Microsoft researchers applied an intersectional approach to test three commercial gender classification algorithms (Buolamwini & Gebru, 2018). The researchers provided skin type annotations for unique subjects in two datasets and built a new facial image dataset that is balanced by gender and skin type (Buolamwini & Gebru, 2018). Analysis of the dataset benchmarks revealed that all three algorithms performed the worst on darker-skinned females, with error rates up to 34.7% higher than for lighter-skinned males (Buolamwini & Gebru, 2018). The classifiers also performed more effectively on male faces (Buolamwini & Gebru, 2018). The researchers suggested that darker skin may not be the only factor responsible for misclassification and that darker skin may instead be highly correlated with facial geometrics or gender presentation standards (Buolamwini & Gebru, 2018). Noting that default camera settings are often optimized to better expose lighter skin than darker skin, the researchers concluded that under-and overexposed images lose crucial information making them inaccurate measures of classification within artificial intelligence systems (Buolamwini & Gebru, 2018). The report also emphasizes the need for increased diversity of phenotypic and demographic representation in face datasets and algorithmic evaluations since “[i]nclusive benchmark datasets and subgroup accuracy reports will be necessary to increase transparency and accountability in artificial intelligence” (Buolamwini & Gebru, 2018).

In 2019, the National Institute of Standards and Technology (“NIST”) released a series of reports on ongoing face recognition vendor tests (“FRVT”). Using both one-to-one verification algorithms and one-to-many identification search algorithms submitted to the FRVT by corporate research and development laboratories and a few universities, the NIST Information Technology Laboratory quantified the accuracy of face recognition algorithms for demographic groups defined by sex, age, and race or country of origin (Natl. Inst. of Stand. & Technol. [NIST], 2018). The NIST used these algorithms with four large datasets of photographs collected in U.S. governmental applications (3) (Natl. Inst. of Stand. & Technol. [NIST], 2018), which allowed researchers to process a total of 18.27 million images of 8.49 million people through 189 mostly commercial algorithms from 99 developers (Natl. Inst. of Stand. & Technol. [NIST], 2018).

The FRVT report confirms that a majority of the face recognition algorithms tested exhibited demographic differentials of various magnitudes in both false negative results (rejecting a correct match) and false positive results (matching to the wrong person) (Crumpler, 2020). In regard to false positives, the NIST found: (1) that false positive rates are highest in West and East African and East Asian people, and lowest in Eastern European individuals (Natl. Inst. of Stand. & Technol. [NIST], 2018) (4), (2) that, with respect to a number of algorithms developed in China, this effect is reversed, with low false positives rates on East Asian faces; (3) that, with respect to domestic law enforcement images, the highest false positive rates are in American Indians, with elevated rates in African American and Asian populations; (4) and that false positives are higher in women than men, and this is consistent across algorithms and datasets (Natl. Inst. of Stand. & Technol. [NIST], 2018). In regard to false negatives, the NIST found: (1) that false negatives are higher in Asian and American Indian people in domestic mugshots; (2) that false negatives are generally higher in people born in Africa and the Caribbean, the effect being stronger in older individuals (5) (Natl. Inst. of Stand. & Technol. [NIST], 2018).

Encouragingly, the NIST concluded that the differences between demographic groups were far lower in algorithms that were more accurate overall (Natl. Inst. of Stand. & Technol. [NIST], 2018). This conclusion signals that as FRTs continue to evolve, the effects of bias can be reduced (Crumpler, 2020). Based on its finding that the algorithms developed in the U.S. performed worse on East Asian faces than did those developed in China, the NIST theorized that the Chinese teams likely used training datasets with greater representation of Asian faces, improving their performance on that group (Natl. Inst. of Stand. & Technol. [NIST], 2018). Thus, the selection of training data used to build algorithmic models appears to be the most important factor in reducing bias (Crumpler, 2020).

Although both the “Gender Shades” and FRVT reports identify under-representative training sets as major sources of algorithmic bias, another recent study of commercial facial algorithms led by Mei Wang showed that “[a]ll algorithms . . . perform the best on Caucasian testing subsets, followed by Indians from Asia, and the worst on Asians and Africans. This is because the learned representations predominately trained on Caucasians will discard useful information for discerning non-Caucasian faces” (Wang, 2019). Furthermore, “[e]ven with balanced training, we see that non-Caucasians still perform more poorly than Caucasians. The reason may be that faces of coloured skins are more difficult to extract and pre-process feature information, especially in dark situations” (Wang, 2019).

Between 2014 and 2018, the accuracy of facial recognition technology has increased 20-fold (Natl. Inst. of Stand. & Technol. [NIST], 2018). However, further applications of FRT will almost certainly bring new challenges if the prevalence of bias remains unchecked. According to Jan Lunter, co-founder and CEO of Innovatrics, facial recognition companies can approach the issue of bias using the insights that the biometrics industry has gained over the past two decades. “Any failure to use these techniques,” Lunter warns, “will not only fan public mistrust, but also inhibit the iterative pace of improvement shown over the past five years” (Natl. Inst. of Stand. & Technol. [NIST], 2018).

II. Current State and Federal Regulatory Schemes

Against a backdrop of scant federal regulation of commercial AI use, including FRTs, several states have adopted their own regulatory schemes to govern the emergent technology. Illinois (740 Ill Comp. Stat), Washington (Wash. Rev. Code), California (Cal. Civ. Code), and Texas (11 Tex. Bus. & Com. Code) have each enacted legislation that targets private sector use of biometric information, including facial images. The states’ legislative schemes commonly define biometric identifiers that encompass facial images by describing them as “face geometry” or unique biological patterns that identify a person (Yeung et al, 2020). However, the states each employ vastly different methods of enforcement. In Texas and Washington, only the state attorney general has enforcement power (11 Tex. Bus. & Com. Code). In California, the state attorney general and the consumer share responsibility for taking action against entities that violate privacy protections (Cal. Civ. Code). While, in Illinois, any person has the right to pursue action against firms and obtain damages between $1,000 and $5,000 per violation (740 Ill. Comp. Stat). Consequently, companies such as Google, Shutterfly, and Facebook have been sued in Illinois for collecting and tagging consumers’ facial information (Yeung et al., 2020).

Facial recognition bans, which range in scope, are on the rise at the municipal level. In September 2020, Portland, Oregon, banned facial recognition use by both public and private entities, including in places of “public accommodation,” such as restaurants, retail stores and public gathering spaces (Metz, 2020). The Portland, Oregon ban does allow private entities’ use of FRTs (1) to the extent necessary to comply with federal, state, or local laws; (2) for user verification purposes to access the user’s own personal or employer-used communication and electronic devices; or (3) in automatic face detection services in social media apps (Hunton Andrews Kurth LLP, 2020). Similarly, Portland, Maine passed an ordinance in November 2020 banning both the city and its departments and officials from “using or authorizing the use of any facial surveillance software on any groups or members of the public” (Heater, 2020). The ordinance allows members of the public to sue if “facial surveillance data is illegally gathered and/or used” (Heater, 2020). Importantly, the Portland, Maine ban does not apply to private companies.

The federal government’s national AI strategy continues to take shape with constant new developments. On November 17, 2020, the Director of the Office of Management and Budget (“OMB”), pursuant to Executive Order 13859, issued a memorandum addressed to the heads of executive departments and agencies that provided guidance for the regulation of non-governmental applications of “narrow” or “weak” AI (6) (The White House, 2020). The OMB’s memo briefly recognized the potential issues of bias and discrimination in AI applications and recommended that agencies “consider in a transparent manner the impacts that AI applications may have on discrimination.” Specifically, the OMB recommended that when considering regulatory or non-regulatory approaches related to AI applications, “agencies should consider, in accordance with law, issues of fairness and non-discrimination with respect to outcomes and decision produced by the AI application at issue, as well as whether the AI application at issue may reduce levels of unlawful, unfair, or otherwise unintended discrimination as compared to existing processes.”

Pursuant to the National AI Initiative Act of 2020 (The White House, 2020), the Director of the Office of Science and Technology Policy (“OSTP”) formally established the National AI Initiative Office (the “Office”) on January 12, 2021. The Office is responsible for overseeing and implementing a national AI strategy and acting as a central hub for coordination and collaboration for federal agencies and outside stakeholders across government, industry and academia in AI research and policymaking (The White House, 2020). On October 4, 2022, the OSTP released the Blueprint for an AI Bill of Rights (the “Blueprint”), which “identified five principles that should guide the design, use, and deployment of automated systems to protect the American public in the age of artificial intelligence” (The White House Office of Science and Tech. Policy [OSTP], 2022a).

The five guiding principles are: 1. Safe and Effective Systems; 2. Algorithmic Discrimination Protections; 3. Data Privacy; 4. Notice and Explanation; and 5. Human Alternatives, Consideration, and Fallback.” (The White House Office of Science and Tech. Policy [OSTP], 2022a). The AI Bill of Rights further provides recommendations for designers, developers, and deployers of automated systems to put these guiding principles into practice for more equitable systems. The Biden-Harris administration has also announced progress across the Federal government that has advanced the Blueprint’s guiding principles, including actions from the Department of Labor, the Equal Employment Opportunity Commission, the Consumer Financial Protection Bureau, and the Federal Trade Commission (“FTC”) (The White House Office of Science and Tech. Policy [OSTP], 2022b).

Most recently, U.S. Senate Majority Leader Charles Schumer has spearheaded efforts to manage AI by circulating a framework that outlines a proposed regulatory regime for AI technologies. Schumer declared on the Senate floor, “Congress must move quickly. Many AI experts have pointed out that the government must have a role in how this technology enters our lives. Even leaders of the industry say they welcome regulation.” Schumer’s nod towards industry leaders is likely in reference to the several congressional panels that held hearings on AI with industry experts during the week of May 16, 2023. Most notably, Sam Altman, the CEO of OpenAI, the company known for promulgating ChatGPT, testified before a Senate committee on May 16, 2023, imploring legislators to regulate the fast-growing AI industry. Altman proposed a three-point plan for regulation that called for: 1. A new government agency with AI licensing authority, 2. The creation of safety standards and evaluations, and 3. Required independent audits. In response to Altman’s plea, Senator Schumer met with a group of bipartisan legislators to begin drafting comprehensive legislation for AI regulation.

The FTC has already taken an active role in regulating private sector development and use of FRT, as evidenced by its recent settlements with Facebook and Everalbum. Further solidifying the FTC’s regulatory stance, acting FTC Chairwoman Rebecca Kelly Slaughter made remarks at the Future of Privacy Forum specifically tying the FTC’s role in addressing systemic racism to the digital divide, AI and algorithmic decision-making, and FRTs (Federal Trade Commission [FTC], 2019).

On April 19, 2021, the FTC published a blog post announcing the Commission’s intent to bring enforcement actions related to “biased algorithms” under section 5 of the FTC Act, the Fair Credit Reporting Act, and the Equal Credit Opportunity Act (Federal Trade Commission [FTC], 2021). Importantly, the statement expressly notes that, “the sale or use of—for example—racially biased algorithms” falls within the scope of the FTC’s prohibition of unfair or deceptive business practices (Federal Trade Commission [FTC], 2021). The FTC also provides guidance on how companies can “do more good than harm” in developing and using AI algorithms by auditing its training data and, if necessary, “limit[ing] where or how [they] use the model;” testing its algorithms for improper bias before and during deployment; employing transparency frameworks and independent standards; and being transparent with consumers and seeking appropriate consent to use consumer data (Federal Trade Commission [FTC], 2021).

III. Argument

The fledgling federal, state, and municipal AI and FRT regulations exist in a loose patchwork that will likely complicate enforcement and compliance for private companies. These complications could hamper or, in some cases, de-incentivize the reduction of bias in FRTs as companies could seek shelter in whichever jurisdiction is most permissive. The federal government’s creation of the National AI Initiative Office does not ensure reductions in FRT bias because the Office is primarily authorized to facilitate AI innovation and cooperation between the government and private companies, rather than addressing any inherent biases present in the FRTs. The FTC’s recent enforcements against private use of FRTs and its recent guidelines indicate that it has an interest in addressing the use of FRTs and FRT bias. However, the FTC possesses limited authority in this context and has historically struggled to compel compliance from large corporations. Thus, a new agency, specifically authorized to regulate and eliminate issues of bias that arise from commercial FRT applications, is needed to effectively address the presence and effect of bias within FRTs.

A. The States’ privacy protections for consumers, comprised of a patchwork of state and municipal regula tions, are inadequate to sufficiently address the issues of bias anticipated from the commercial use of FRTs.

In the absence of federal laws that regulate the commercial use of AI, much less FRTs, state and city laws have attempted to fill the regulatory gap. State governments may be regarded and valued as “living laboratories” in some respects, but their collective piecemeal legislation concerning commercial AI and FRT use may negatively impact the reduction of bias in FRTs and could likely lead to a deregulatory “race to the bottom.”

Significantly, three states–Illinois, Texas and Washington—have recognized the urgent need to address the burgeoning use of AI in the private sector and put privacy protections in place for consumers. The Illinois Biometric Information Privacy Act, passed in 2008, requires commercial entities to obtain written consent in order to capture an individual’s biometric identifiers (including face geometry) or sell or disclose a person’s biometric identifier (740 Ill. Comp. Stat). The Illinois Act also places security and retention requirements on any collected biometric data (740 Ill. Comp. Stat). Although Texas and Washington have enacted similar laws, their laws vary significantly from Illinois’ in that only the attorney generals are authorized to enforce the laws against commercial entities (11 Tex. Bus. & Com. Code). Illinois’ law, on the other hand, includes a private right of action, which has led to several lawsuits against companies such as Clearview AI, Google, and Facebook (Greenberg, 2020; Yeung et al, 2020).

The variance among the entities empowered to enforce these states’ laws will likely create enforcement and compliance difficulties, particularly as it pertains to bias, because AI and FRTs inherently transcend state borders. Based on recent studies of the presence of bias in commercial FRTs, the selection of training data used to build algorithmic models appears to be the most important factor in reducing bias (Crumpler, 2020). Thus, the reduction of bias in commercial FRTs would be significantly hindered if companies are unsure whether they have access to certain images based on specific state laws. For example, Everalbum’s settlement with the FTC revealed that the international company compiled FRT training datasets by combining facial images it had extracted from Ever users’ photos with facial images obtained from publicly available datasets (Everalbum, Inc., n.d.). Everalbum’s FRT development was geographically constrained on a state-by-state basis to exclude images from users believed to be residents of Illinois, Texas, Washington, or the European Union (Everalbum, Inc., n.d.). From the perspective of increasing representative training datasets, the company’s exclusion of facial images from users in Texas and Illinois, specifically, would have negatively impacted the representation of Latinx people and other racial minorities (7) (Krogstad, 2020).

Given uncertainty among AI and FRT developers within the patchwork state regulatory scheme, paired with researchers’ recommendations to increase phenotypic and demographic representation in face datasets and algorithmic evaluations (Buolamwini & Gebru, 2018), companies will likely want to conduct business in locations that enable them to have access to large amounts of data. In response, states may avoid enacting AI and FRT regulations that deter companies from conducting business in those states, resulting in what is termed as a deregulatory “race to the bottom” (Chen, 2022). If a “race to the bottom” situation was to occur in response to the patchwork of state AI regulations, then companies would likely seek to build and train FRTs in those states where consumers had less rights to their biometric data since the companies would have access to more information to compile larger datasets.

On the one hand, enabling companies’ ability to compile larger datasets seems like a great avenue to reduce bias in FRT applications, as the larger datasets would provide increased phenotypic and demographic diversity. However, a lack of state standards governing the quality and collection of biometric data could negatively impact FRT accuracy and, in turn, exacerbate the presence of biased results. According to one study, non-Caucasians may perform more poorly than Caucasians on FRTs, even with balanced training, because “faces of coloured skins are more difficult to extract and pre-process feature information, especially in dark situations” (Wang et al., 2019). Similarly, the “Gender Shade” researchers noted that default camera settings are often optimized to better expose lighter skin than darker skin (Buolamwini & Gebru, 2018). This observation led the researchers to conclude that under-and overexposed images lose crucial information making them inaccurate measures of classification within artificial intelligence systems (Buolamwini & Gebru, 2018). If biased FRT performance is linked to the difficulty of extracting and pre-processing feature information from non-Caucasian faces, especially in dark situations; and, if sub-optimal camera lightening of non-Caucasian faces often produces images that lack crucial information rendering them inaccurate datapoints; then, lax state regulations on the quality and collection of biometric data will likely widen the discrepancy between FRTs’ performance on Caucasian and non-Caucasian faces, undermining efforts to reduce bias in commercial FRT use.

B. The current federal regulatory scheme lacks the scope and capacity to sufficiently address the issues of bias anticipated from the commercial use of FRTs.

The U.S. federal government, in passing the National AI Initiative Act of 2020 and creating the National AI Initiative Office (the “Office”), decided to primarily focus its resources on the support and growth of AI and its attendant technologies, including FRTs (Gibson, Dunn & Crutcher LLP, 2021). The Act also (1) expanded and made permanent the Select Committee on AI, which will serve as the senior interagency body responsible for overseeing the National AI Initiative; (2) codified the National AI Research Institutes and the National Sciences Foundation, collaborative institutes that will focus on a range of AI research and development areas, into law; (3) expanded AI technical standards to include an AI risk assessment framework; and (4) codified an annual AI budget rollup of Federal AI research and development investments (The White House Office of Science and Tech. Policy [OSTP], 2021). Further, on January 27, 2021, President Biden signed a memorandum titled, “Restoring trust in government through science and integrity and evidence-based policy making,” setting in motion a broad review of federal scientific integrity policies and directing agencies to bolster their efforts to support evidence-based decisions making (The White House Office of Science and Tech. Policy [OSTP], 2021). In spite of these nascent attempts to federally regulate commercial use of FRTs, the existing commercial applications of FRTs and the instances of bias that arise from such use remain largely unregulated.

The National AI Initiative Office lacks the capacity and authority to regulate bias arising from current commercial FRT use since, according to its enabling statute, the Office is principally concerned with supporting public and private AI innovation. The National AI Initiative Act describes the Office’s responsibilities as serving as a liaison between the government, industry, and academia; outreaching to the public, and promoting innovation (The White House, 2020).

None of the enumerated responsibilities described in the National AI Initiative Act authorize the Office to specifically regulate existing commercial AI use, let alone address any issues of bias. The first two responsibilities establish the Office’s authority to “provide technical and administrative support” to other federal AI Initiative committees and serve as a liaison on federal AI activities between a broadly defined group of public and private entities. The last two responsibilities charge the Office with reaching out to “diverse stakeholders” and promoting interagency access to the AI Initiative’s activities. The Office’s enabling statute does not clearly indicate whether the regulatory body has enforcement authority on private actors as there is no provision that confers on the Office the ability to promulgate rules or regulations. Likewise, the Office does not seem to have the power to impose sanctions in order to ensure industry compliance. Instead, the Office is focused on building coordination between the private sector and governmental entities to promote further AI innovation. Thus, the Office does not have explicit regulatory authority over any existing private use of AI or FRTs.

Supporters of the National AI Initiative Act and the Office may argue that the Office is appropriately situated to address issues of bias arising from the commercial use of FRTs, however that argument is undermined by the express statutory language of the Act. A supporter of the Office may point to the entity’s responsibility to serve as a liaison on federal AI activities between public and private entities to argue that, by facilitating the exchange of technical and programmatic information that could address bias in AI, the Office would help FRT developers reduce bias. However, the statute does not appear to enable the Office to influence or contribute to the substantive contents of the information shared between the public and private sectors about the AI Initiative activities. If the Office lacks the ability to influence the substance of information exchanged, then it also lacks the ability to specifically direct information sharing that could redress bias in commercial AI applications. A supporter of the Office may also point to its outreach responsibility to argue that the Office will work to address bias by reaching out to diverse stakeholders, including civil rights and disability rights organizations. Yet, the statutory language is vague as to the substance of this “regular public outreach” responsibility. Without a clearer indication that the Office’s public outreach efforts are directed toward or will somehow result in a reduction in AI and FRT bias, the assumption that coordinating public outreach with diverse stakeholders will sufficiently address bias in commercial FRT use remains unfounded. Hence, reducing bias that arises from the commercial use of FRTs is not an articulated central focus, nor an explicitly intended effect, of the Office’s enabling statute.

Close analysis of the statutory language establishing the National AI Initiative and the Office reveals that the Office will likely operate more like a governmental think-tank to ensure coordinated AI innovation than a regulatory body with enforcement power. Such a scheme is inadequate to properly address the existing issues of bias shown in today’s commercial FRTs since the AI Initiative will likely promulgate industry standards that stem from and reflect the market itself, including its apparent biases.

The few FTC regulatory decisions that have been handed down concerning existing commercial FRT applications are products of the FTC’s recent actions to regulate private AI use (Facebook, Inc., n.d.). Based on its latest posts and statements, the FTC anticipates broadening its regulation of private AI and FRT use to not only focus on user consent, but also biased algorithms (Jillson, 2021). However, the FTC has limited enforcement power to sufficiently address the wide-ranging applications of FRTs and reduce the perpetuation of bias.

The Federal Trade Commission Act empowers the FTC to, among other things:

(a) prevent unfair methods of competition and unfair or deceptive acts or practices in or affecting commerce;
(b) seek monetary redress and other relief for conduct injurious to consumers; and
(c) prescribe rules defining with specificity acts or practices that are unfair or deceptive, and establishing requirements designed to prevent such acts or practices (15 U.S.C. §§ 41-58).

As stated in its enabling statute, the FTC’s enforcement power is limited to “unfair or deceptive acts or practices in or affecting commerce” (15 U.S.C. §§ 41-58) The FTC asserts its authority over certain issues or subject areas by deeming a certain commercial practice unfair or deceptive, which is exactly what the FTC did when it released its recent AI blog post categorizing the use or sell of “biased algorithms” as an unfair and deceptive practice. Yet, the FTC will likely run into future enforcement issues in trying to prevent the use and sale of biased algorithms because they lack the willingness to enforce orders and expertise in AI training and development. Despite the FTC’s recent blog post indicating its intention to bring enforcement actions related to biased algorithms, FTC Commissioner Rohit Chopra provided a statement to the Senate noting that “Congress and the Commission must implement major changes when it comes to stopping repeat offenders” and that “since the Commission has shown it often lacks the will to enforce agency orders, Congress should allow victims and state attorneys general to seek injunctive relief in court to halt violations of FTC orders (Federal Trade Commission [FTC], 2021).

In support of his first suggestion concerning the issue of repeat offenders, Commissioner Chopra emphasized that, “[w]hile the FTC is quick to bring down the hammer on small businesses, companies like Google know that the FTC simply is not serious about holding them accountable” (Federal Trade Commission [FTC], 2021). If the FTC is currently struggling to “turn the page on [their] perceived powerlessness” (Federal Trade Commission [FTC], 2021), then it follows that it is most likely ill-suited to successfully take on emerging global leaders in commercial AI technology. Furthermore, the Commissioner’s plea for Congress to allow victims and state attorneys general to access the courts for injunctive relief underscores the FTC’s inability and unwillingness to enforce its orders. Shifting the burden onto consumers and judges to regulate the exploding commercial use of FRTs and reduce bias is less than ideal as the courts lack the expertise and resources to adequately address bias in commercial FRT use. Also, courts are bound by justiciability principles, which limits their ability to regulate and reduce bias. Therefore, Congress should create a new agency that is solely authorized to address issues of bias in commercial FRT use, has power to regulate, and teeth to go after private parties who violate its regulations.

C. Congress must establish a new federal agency specifically, but not solely, authorized to regulate and eliminate issues of bias that arise from commercial FRT applications.

In order to effectively address the pervasiveness of bias in private FRT use, Congress must establish a new regulatory agency specifically, but not solely, authorized to regulate and eliminate issues of bias that arise from commercial FRT applications. The new agency should be created according to the following enabling statute to ensure its appropriate scope and capacity:

The [agency] is empowered, among other things, to:

(a) prevent private entities’ development, use, or sale of FRTs in circumstances that perpetuate bias based on ethnic, racial, gender, and other human characteristics recognizable by computer systems;
(b) seek monetary redress or other relief for injuries resulting from the presence of bias in FRTs; (c) prescribe rules and regulations defining with specificity circumstances known or reasonably foreseeable to perpetuate bias that is prejudicial to established human and legal rights, and establishing standards designed to prevent such circumstances;
(d) gather and compile data and conduct investigations related to private entities’ development, testing, and application of FRTs; and
(e) make reports and legislative recommendations to Congress and the public. (8)

Part (a) of the new agency’s enabling statute delineates the scope of the agency’s enforcement power to specifically regulate private entities’ development, use or sale of FRTs in settings that perpetuate bias. The phrase “development, use, or sale” is designed to extend the agency’s regulatory scope to include the development or creation of FRTs in recognition of the fact that biases can originate from either the algorithm or the training dataset. Including the development stage within the agency’s regulatory authority will allow the agency to effectively regulate the sources of bias—the algorithm, training datasets, and photo quality. Additionally, the inclusion of all three stages—development, use, and sale—enable the agency to have the conceptual framework and authority to regulate any future sources of bias that are yet to be discovered (9) (Learned-Miller et al., 2020).

Part (b) confers the agency the power to impose sanctions in the form of monetary penalties or other appropriate type of relief for injuries caused by a private party’s violation of the agency’s regulations. Part (b) is of utmost importance since it will give the agency power to bring down the hammer on violating entities and shirk the perception of “powerlessness.” The agency will compel compliance from large companies by bringing timely actions against violating parties, requiring violating parties to make material changes to their algorithms that eliminate or significantly reduce bias, and maintaining a reputation for rigorously holding companies accountable for their algorithms.

Part (c) functions hand-in-hand with Part (b) in that the agency’s promulgation of rules and regulations creates the legal claims through which the agency can seek monetary redress or other forms of relief from violating companies. Requiring the agency to prescribe rules and regulations that specifically define circumstances known or reasonably foreseeable to perpetuate bias will require significant technical expertise. The agency should employ and regularly consult with preeminent AI and FRT scholars and researchers so that it can stay abreast of industry standards, norms, and developments. The agency must also develop rigorous testing standards to identify and address algorithms’ rates of bias, which will require it to compile large datasets that are phenotypically and demographically representative.

Part (d) significantly empowers the agency to continually request information from private FRT developers so that it can promulgate rules and standards that can effectively address the identified sources of bias in commercial FRT applications. Without the power to gather and compile data, the agency’s regulations and standards would run the risk of becoming obsolete or irrelevant to the FRT industry, which would hinder its ability to reduce bias. Similarly, the power to conduct investigations related to FRT development, testing, and applications is crucial to the agency’s regulatory authority so that the agency can actively ensure companies’ compliance without needing to wait on injured parties, who often lack AI expertise or access to representation, to bring claims. Based on its investigations, the agency can further ensure the sustained reduction in FRT bias by making reports and recommendations to Congress and the public.

Part (e) can be best realized by the agency because of its broad authority to regulate every aspect of FRT development and application. Thus, the agency sits at a critical juncture between FRT developers, legislators, and the public. Consequently, the agency can emphasize legislative reform as needed to effectively reduce bias and contribute to a nascent body of knowledge that the public has only begun to understand.

A new federal agency, empowered to investigate and regulate FRT development, testing, and application can reduce the presence of bias more effectively than the current regulatory scheme because of its broad authority and enforcement power. The FTC is limited in its authority to regulate bias, and its regulatory power has repeatedly bowed to the will of large corporations. Furthermore, it is not clear whether the Office has the authority to even promulgate rules or standards. Yet, FRT technology is a growing market, and researchers have only scratched the surface of how FRTs perpetuate bias. To this end, the Association of Computing Machinery’s U.S. Technology Policy Committee observed that industry and government have adopted FRTs “ahead of the development of principles and regulations to reliably assure their consistently appropriate and non-prejudicial use” (Association for Computing Machinery [ACM], 2020). A new agency, specifically targeting the development, training, and application of FRTs can have the necessary breadth and expertise to reduce existing sources of bias and discover unknown sources of bias. Furthermore, the agency’s narrowly tailored focus on FRTs can help to lay a foundation for its future expanded regulatory authority over additional AI attendant technologies, which are likely more complex systems. Since large corporations have not dealt with the new agency yet, the agency will be able to set itself apart from agencies with waning respect from corporations by strictly enforcing its regulations, erring on the side of caution, and crafting settlement agreements with provisions that require violators to make material changes to reduce FRT bias. Though the existence of completely unbiased FRTs is sure to be difficult to realize, the new agency will deploy all of its authority and resources to reducing FRT bias to the point of elimination.


The inundation of commercial facial recognition technology coupled with a lagging federal regulatory framework to govern commercial FRT development and use has led to a precarious environment where individuals bear the un- due burden of redressing unprecedented harms. The following policy recommendations, while ambitious, aim to support a national regulatory scheme that would reduce the frequency and severity of FRT bias and discrimination:

  • Establish a federal agency with the explicit authority to regulate commercial AI and its attendant technolo- gies, like FRTs, in accordance with the following enabling statute:

The [agency] is empowered, among other things, to:

(a) prevent private entities’ development, use, or sale of FRTs in circumstances that perpetuate bias based on ethnic, racial, gender, and other human characteristics recognizable by computer systems;
(b) seek monetary redress or other relief for injuries resulting from the presence of bias in FRTs; (c) prescribe rules and regulations defining with specificity circumstances known or reasonably foreseeable to perpetuate bias that is prejudicial to established human and legal rights, and establishing standards designed to prevent such circumstances;
(d) gather and compile data and conduct investigations related to private entities’ development, testing, and application of FRTs; and
(e) make reports and legislative recommendations to Congress and the public.

  • Create uniform guidelines for states’ regulation of the commercial collection and use of biometric data;
  • Develop and encourage the increased implementation of phenotypically and demographically diverse face datasets in commercial FRT development, training, and evaluation.


  1. Khortlan Becton graduated summa cum laude from the University of Alabama with Bachelor of Arts degrees in Religious Studies and African American Studies, received a Master of Theological Studies from Vanderbilt Divinity School, and a Juris Doctor from Temple School of Law. Khortlan is a Truman Scholar Finalist, a member of Phi Beta Kap- pa, and recipient of numerous awards from the various academic institutions that she has attended.
    While attending law school, Khortlan began studying and advocating for the regulation of artificial intelligence technologies, including facial recognition technology. She served as lead author for a summary of recent literature on algorithmic bias in decision-making and related legal implications. That paper’s co-author, Professor Erika Douglas (Temple School of Law), presented the literature review at the American Bar Association’s Antitrust Spring Meeting in 2022.Through holding various service and leadership roles, Becton has developed a deep appreciation for creative and collaborative problem-solving to address intergenerational issues of poverty and systemic inequality. Becton has continued to pursue her passion for education and justice by launching The Restorative Education Institute. The Institute is a non-profit organization purposed to equip youth and adults to practice anti-racism through historical education and substantive reflection.
  2. The FTC, which is playing an active role in the misuse of facial recognition, previously imposed a $5 billion penalty and new privacy restrictions on Facebook in 2019. Similar to the allegations against Everalbum, the complaint against Facebook alleged that Facebook’s data policy was deceptive to users who have Facebook’s facial recognition setting because that setting was turned on by default, while the updated data policy suggested that users would need to opt-in to having facial recognition enabled.
  3. The four large datasets of photographs are: (1) Domestic mugshots collected in the U.S.; (2) Application photographs from a global population of applicants for Immigration benefits; (3) Visa photographs submitted in support of visa applicants; and (4) Border crossing photographs of travelers entering the U.S.
  4. This effect is generally large, with a factor of 100 more false positives between countries.
  5. These differing results relate to image quality: The mugshots were collected with a photographic setup specifically standardized to produce high-quality images across races; the border crossing images deviate from face image quality standards.
  6. The OMB memorandum defines “narrow” AI as “go[ing] beyond advanced conventional computing to learn and perform domain-specific or specialized tasks by extracting information from data sets, or other structured or unstructured sources of information.”
  7. According to Pew Research, Texas is one of two states with the most Latinx people at 11.5 million. Illinois’ Latinx population increased from 2010 to 2019 by 185,000 people.
  8. The new agency’s enabling statute is modeled after the Federal Trade Commission Act because the Act succinctly embodies the power of a narrowly focused agency. The FTC Act is primarily focused on “unfair and deceptive acts or practices affecting commerce,” which has contributed to the FTC’s broad authority. The new agency will need a similar breadth in their jurisdictional scope since researchers have only begun to scratch the surface of bias in FRT applications.
  9. Specifically defining the concepts used to describe the creation and management of FRTs is of utmost importance to delineating the scope of not only the AI attendant tech- nology, but also the breadth of an agency’s regulatory framework. Researchers have recently endeavored to provide specific definitions for the creation of a federal regulatory scheme for FRTs that will likely be a necessary addition to the enabling statutory language proposed here.


11 Tex. Bus. & Com. Code § 503.001 (2019). 11 Tex. Bus. & Com. Code § 503.001(D).
15 U.S.C. §§ 41-58.

740 Ill. Comp. Stat. 14/1 (2008).

740 Ill. Comp. Stat. 14/20 (2008).

Amazon Web Services [AWS]. (2021). The Facts on Facial Recognition with Artificial Intelligence. Retrieved from

Association for Computing Machinery [ACM]. (2020, June 30). U.S. Technology Policy Committee: Statement on Principles and Prerequisites for the Development, Evaluation and Use of Unbiased Facial Recognition Technologies. Retrieved from recognition-tech-statement.pdf

Brownlee, J. (2019, July 5). A gentle introduction to deep learning for face recognition. Machine Learning Mastery. Retrieved from

Buolamwini, J., & Gebru, T. (2018). Gender shades: Intersectional accuracy disparities in commercial gender classification. Proceedings of Machine Learning Research 81, 1-15.

Cal. Civ. Code § 1798.100 – 1798.199.100 (2018).

Cal. Civ. Code § 1798.150, § 1798.199.90 (2018).

Calvello, M. (2019, Oct 23). 22 eye-opening facial recognition statistics for 2020. Learn Hub. Retrieved from

Chellappa, R., Phillips, P.J., Rosenfeld, A., & Zhao, W. (2003). Face recognition: A literary survey. ACM Journals, 35(4).

Chen, J. (2022, Oct. 3). Race to the bottom. Investopedia. Retrieved from race-bottom.asp#:~:text=The%20race%20to%20the%20bottom%20refers%20to%20a%20competi- tive%20situation,can%20also%20occur%20among%20regions

Crumpler, W. (2020, May 1). The problem of bias in facial recognition. Center for Strategic and International Studies Blog Post. Retrieved from problem-bias-facial-recognition

Everalbum, Inc., (n.d.) FTC Complaint No. 1923172. Retrieved from cases/everalbum_complaint.pdf

Facebook, Inc., FTC Stipulated Order No. 19-cv-2184. Retrieved from documents/cases/182_3109_facebook_order_filed_7-24-19.pdf

Federal Trade Commission [FTC]. (2019, July 24). FTC imposes $5 billion penalty and sweeping new privacy restrictions on Facebook. Retrieved from ftc-imposes-5-billion-penalty-sweeping-new-privacy-restrictions

Federal Trade Commission [FTC]. (2021, April 20). Prepared opening statement of Commissioner Rohit Chopra, U.S. Senate Committee on Commerce, Science, and Transportation hearing on “Strengthening the FederalTrade Commission’s authority to protect consumers”. Retrieved from files/documents/public_statements/1589172/ final_chopra_opening_statement_for_senate_commerce_committee_20210420.pdf

Federal Trade Commission [FTC]. (2021, Feb. 10). Protecting Consumer privacy in a time of crisis, remarks of acting Chairwoman Rebecca Kelly Slaughter, future of privacy forum. Retrieved from system/files/documents/public_statements/1587283/fpf_opening_remarks_210_.pdf

Federal Trade Commission [FTC]. (2021, Jan. 8). Statement of Commissioner Rohit Chopra, in the matter of Everalbum and Paravision, Commission File No. 1923172. Retrieved from files/documents/public_statements/1585858/ updated_final_chopra_statement_on_everalbum_for_circulation.pdf

Gibson, Dunn & Crutcher LLP. (2021, April 23). Artificial intelligence and automated systems legal update (1Q21). Retrieved from artificial-intelligence-and-automated-systems-legal-update-1q21/#_ftn3

Greenberg, P. (2020, Sept. 18). facial recognition gaining measured acceptance. State Legislatures Magazine. Retrieved from facial-recognition-gaining-measured-acceptance-magazine2020.aspx

Grisales, C. (May 18, 2023). Schumer meets with bipartisan group of senators to build a coalition for AI law, schumer-meets-with-bipartisan-group-of-senators-to-build-a-coalition-for-ai-law

Grother, P., Ngan, M., & Hanaoka, K. (20190. Face Recognition Vendor Test (FRVT) Part 3: Demographic Effects. Natl. Inst. of Stand. & Technol., NISTIR 8280. Retrieved from

The White House. (2020, Nov. 17). Guidance for regulation of artificial intelligence applications, H.R. 6395, § 5102. 2, OMB Op. Retrieved from uploads/2020/11/M-21-06.pdf

Heater, B. (2020, Nov. 4). Portland, Maine passes referendum banning facial surveillance, Retrieved from portland-maine-passes-referendum-banning-facial-surveillance/

Hunton Andrews Kurth LLP. (2020, Sept. 10). Portland, Oregon first to ban private-sector use of facial recognition technology. Privacy & Information Security Law Blog. Retrieved from https://www.huntonprivacyblog. com/2020/09/10/portland-oregon-becomes-first-jurisdiction-in-u-s-to-ban-the-commercial-use-of- facial-recognition-technology/

IBM. (2020, June 3). What is artificial intelligence (AI)? IBM Cloud Education. Retrieved from com/cloud/learn/what-is-artificial-intelligence

Jillson, E. (2021, April 19). Aiming for truth, fairness, and equity in your company’s use of AI. Federal Trade Commission [FTC]. Retrieved from aiming-truth-fairness-equity-your-companys-use-ai

Kang, C. OpenAI’s Sam Altman Urges A.I. Regulation in Senate Hearing,, (May 16, 2023), https://

Krogstad, J.M. (2020, July 10). Hispanics have accounted for more than half of total U.S. population growth since 2010. Pew Research. Retrieved from hispanics-have-accounted-for-more-than-half-of-total-u-s-population-growth-since-2010/#:~: text=Some%20of%20the%20nation’s%20largest,of%20the%20nation’s%20Hispanic%20population

Learned-Miller, E., Ordóñez, V., Morgenstern, J., & Buolamwini,J. (2020, May 29). Facial recognition technologies in the wild: A call for a federal office. Center for Integrative Research in Computing and Learning Sciences. Retrieved from

MarketsandMarkets, (2020, Dec. 2). Facial recognition market worth $8.5 billion by 2025. Retrieved from

Metz, R. (2020, Sept. 9). Portland passes broadest facial recognition ban in the US., Retrieved from

Najibi, A. (2020, Oct. 4). Racial discrimination in face recognition technology. Harvard University, Science in the News. Retrieved from racial-discrimination-in-face-recognition-technology/#

National Artificial Intelligence Initiative Act. U.S. Public Law 116-283 (2020).

Natl. Inst. of Stand. & Technol. [NIST]. (2018, Nov. 30). NIST evaluation shows advance in face recognition software’s capabilities. Retrieved from nist-evaluation-shows-advance-face-recognition-softwares-capabilities

Sakin, N. (2021, Feb. 11). Will there be federal facial recognition in the U.S.? International Association of Privacy Professionals (IAPP). Retrieved from

Samuel, S. (2019, March 6). A new study finds a potential risk with self-driving cars: Failure to detect dark-skinned pedestrians. Retrieved from self-driving-car-racial-bias-study-autonomous-vehicle-dark-skin

Solender, A. & Gold, A. (April 13, 2023). Scoop: Schumer lays groundwork for Congress to regulate AI,,

Thales Group. (2021, March 30). Facial recognition: Top 7 trends (tech, vendors, markets, use cases & latest news). Retrieved from biometrics/facial-recognition

The White House Office of Science and Tech. Policy [OSTP]. (2021, Jan. 12). The White House launches the National Artificial Intelligence Initiative Office. Retrieved from briefings-statements/white-house-launches-national-artificial-intelligence-initiative-office/

The White House Office of Science and Tech. Policy [OSTP]. (2022a). Blueprint for an AI bill of rights. Retrieved from

The White House Office of Science and Tech. Policy [OSTP]. (2022b). Biden-Harris administration announces key actions to advance tech accountability and protect the rights of the American public. Retrieved from announces-key-actions-to-advance-tech-accountability-and-protect-the-rights-of-the-american-public/

Towards Data Science. (Sept. 14, 2018). Clearing the confusion: AI vs machine learning vs deep learning differences. Retrieved from vs-deep-learning-differences-fce69b21d5eb

Wang, M., Deng, W., Hu, J., Xunqiang, T., & Huang, Y. (2019). Racial faces in-the-wild: Reducing bias by information maximization adaption network, IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Korea (South). pp. 692-702, doi: 10.1109/ICCV.2019.00078

Wash. Rev. Code § 19.375.020 (2017).

Wash. Rev. Code §19.375.030 (2017).

Yeung, D., Balebako, R., Gutierrez, C.I., & Chaykowsky, M. (2020). Face recognition technologies: Designing systems that protect privacy and prevent bias. Santa Monica, CA: Rand Corporation.

Zakrzewski, C., Tiku, N., Lima C. & Oremus W. (May 16, 2023). OpenAI CEO tells Senate that he fears AI’s potential to manipulate views, Washington, ai-congressional-hearing-chatgpt-sam-altman/

Artificial Intelligence and Automation and Its Potential Impact on Race and Class.

Foreward to the 2023 Hooks Institute Policy Papers

To say the world is in the throes of a technological revolution spearheaded by artificial intelligence (“AI”), and automation, may be one of the most understated observations of this century. While “Fake News” ran rampant on social and other media and influenced the November 2016 presidential election, that election provided ample warning of how media manipulated to mislead can have enormous negative consequences for every segment of life, including personal and employment relationships, national security, elections, media, etc.

However, something is intriguing about AI and automation. It gives us access to a futuristic society allowing us to explore unchartered waters. Bill Gates has argued for years that AI has its proven benefits. Potential uses of AI include creating personalized teaching models for students so that educators can maximize students’ educational experiences (Gates, 2023). “AI can reduce some of the world inequities” (Gates, 2023) through its problem-solving capabilities, enhance worker productivity, and “[a]s computer power gets cheaper, GPT’s ability to express ideas will increasingly be like having a white-collar worker available to help you with various tasks” (Gates, 2023).

As for the immediate future, AI may create as many casualties as opportunities. Undergirding the Writers Guild of America, strike were Hollywood writers’ concerns that AI, specifically the program ChatGPT (which can produce creative writing and audio in response to prompts), might reduce or eliminate the need for screenwriters in the future (Fortune 2023).

Individuals, governments, and organizations have used AI in insidious ways. In public housing complexes, surveillance cameras create over-policing of people of color. Despite the lack of evidence showing that Facial Recognition Technology (FRT) makes public housing complexes safer, “many of the 1.6 million Americans who live [in public housing] . . . are overwhelming people of color [who are subjected] to round-the-clock surveillance” (MacMillian, 2023). For example, in the small town of Rolette, North Dakota, the public housing complex has 100 residents un- der the surveillance of 107 cameras, “a number of cameras per capita approaching that found in New York’s Riker Island complex” (MacMillian, 2023).

FRT has led to evictions for minor or alleged infractions that have uprooted lives. In Steubenville, Ohio, a resident was evicted for removing a laundry basket from the washing room of the complex, and another was threatened with eviction because she loaned her key fob to an authorized guest (MacMillian, 2023). The latter resident demonstrated that her vision loss required the help of her friend, who brought her groceries, thus successfully pleading

her case against eviction (MacMillian, 2023). A single mother of two in New Bedford, Massachusetts, who received an eviction notice in 2021, stated that the public housing authority “made [her] life hell” when they alleged that
her ex-husband – who was taking care of their children while his former wife worked during the day and attended school at night – was staying in the apartment without contributing rent in violation of the rules (MacMillian, 2023). Even Bill Gates acknowledges that the new frontier of AI is not without rugged and scorched terrain that produces inequities. Gates recognizes that “market forces won’t naturally produce AI products and services that help the poor- est. The opposite is more likely.” He contends that “[w]ith reliable funding and the right policies, governments and philanthropy can ensure that AIs are used to reduce inequity” (Gates, 2023).


The speed with which technology and automation are transforming the landscape is taking place with unprecedented velocity, even outpacing the rate with which changes occurred during the industrial revolution. “The speed of current breakthroughs has no historical precedent. Compared with previous industrial revolutions, the [technological revolution] is evolving at an exponential rather than a linear pace. Moreover, it is disrupting almost every industry in every country. The breadth and depth of these changes herald the transformation of entire systems of production, management, and governance (Schwab, 2015).

This technological revolution has the potential to raise global income levels and improve the quality of life for populations around the world. To date, those who have gained the most from it have been consumers able to afford and access the digital world (Schwab, 2015). By contrast, African Americans, Hispanics, and marginalized people clustered in service,

warehouse, and other low skills occupations are the least likely beneficiaries of AI and automation gains because they are the most susceptible to job loss because of it. (McFerren & Delavega, 2018).

As the nation and world grapple with the societal impact of AI and Automation, the Hooks Institute remains focused on a core question central to promoting justice and equality: what policies and practices will prevent AI and automation from discriminating against people of color and other marginalized groups? How can AI and automation aid our nation in eliminating racial, economic, health, educational, and other disparities?

The policy papers in this edition analyze the impact of AI and automation in three crucial areas. Khortlan Becton, JD, MTS, explores the urgent need to regulate AI to eradicate existing and potential policies and practices that disproportionately discriminate against African Americans and minorities. Becton proposes the creation of a new federal agency to regulate AI.

Susan Elswick, EdD, LCSW, a faculty member at the University of Memphis School of Social Work, seeks a path to using AI and Automation to provide social work counseling to those in need. Elswick not only explores how effective client counseling is dependent upon access and ability to use technology by clients but also argues that social workers require formal training from institutions of higher learning on how to use AI and automation to benefit their clients.

Meka Egwuekwe, MS, founder and executive director of Code Crew, approaches AI and automation from the perspective of a practitioner who teaches others to write computer code. Recognizing that the world is experiencing a revolution in how work is performed, Egwuekwe proposes recommendations that reskill or upskill the workforce, increased support for startups and small businesses, and a societal framework that will embrace universal basic income as a resource to aid those displaced by AI and Automation.

The world has entered the frontier of AI and Automation. Let’s ensure everyone has an equitable opportunity for life, liberty, and the pursuit of happiness as we embark on this evolving and transformative frontier.

Daphene R McFerren, JD
Executive Director, Benjamin L. Hooks Institute for Social Change

Elena Delavega, PhD
Professor, Department of Social Work

Daniel Kiel, JD
FedEx Professor of Law, Cecil C. Humphreys School of Law Editors

June 30, 2023

Read the full 2023 Policy Papers through this link.


Cole, J. (2023, May 5). Chat GPT is the ‘terrifying’ subtext of the writers’ strike that is reshaping Hollywood. Fortune. Retrieved from hollywood-writers-strike-wga-chatgpt-ai-terrifying-replace-workers/

Gates, B. (2023, March 21). The age of AI has begun: Artificial Intelligence is as revolutionary as mobile phones and the internet. GatesNotes. Retrieved from

MacMillian, Douglas (2023, May 16). Eyes on the poor: Cameras, facial recognition watch over public housing. The Washington Post. Retrieved from surveillance-cameras-public-housing/

McFerren, D. & Delavega, E. (2018) The robots are ready! Are we? Automation, Race and the Workforce.
Hooks Policy Papers. Retrieved from

Schwab, K. (2015, December 12). The fourth industrial revolution: What it means and how to respond. Foreign Affairs. Retrieved from gy=1116078

World Economic Forum. (May 2023). The future of jobs report. Retrieved from WEF_Future_of_Jobs_2023.pdf

Rhyming Colors: A Poem by Mollie A. Steward

A peom by Mollie A. Steward

Once upon a time
In a town that did not rhyme
Lived a girl who wanted playtime fun
But was left alone by everyone

Her mother saw her face so long
And asked her daughter what was wrong
Our girl sobbed, “I’m not asked to play
“And one boy called me ‘freak’ today.”

Mom came near, teardrops to dry
In hopes the sadness soon would fly
By a promise someday all would see
The gem she knew our girl to be

She led her daughter down the hall
To look in a mirror which was full-length tall
“Look,” said Mom, “then smile and say
“I’m beautiful in every way.”

Daughter noted hair of light brown curls
Then said, “My skin covering is purple swirls.”
“Of course,” said Mom, “My own is stripes of pink
“But beauty is deeper than you think.”

So our girl took her cat to sit outside
In an effort not to hide
Another girl walked by as they rested on a mat,
Waved at them and called, “Nice cat!”

“Thanks,” said our girl, “She’s Bella; I’m Abby
“Bella’s a real multi-colored tabby.”
“Want to come see her jump through a hoop?”
“Sure,” said the new friend, approaching Abby’s stoop

To the door stepped two girls, one cat striped gold and pink,
While through the window, Abby saw her mother wink
There just might be some hope that with a bit of time
Things might eventually start to rhyme.

Mollie A. Steward is a retired Professor of Mathematics from Southern New Jersey. Proudly multiracial, she is the daughter of an educator and a brick mason and notes her family has always valued education. She loves writing poetry, and frequent themes are inclusiveness and unity; the concept for this poem grew from musings on those topics.

Learning of the Hooks Institute after watching the Chicago Stories special on Ida B. Wells, Steward recalled a memory of a meeting with Dr. Benjamin L. Hooks at the Convention of the New Jersey Education Association some years ago. “In a personal exchange, he graciously shared some hopeful scripture, a portion of Romans 5:20 – ‘…But where sin abounded, grace did much more abound.’ (KJV),” says Steward. Steward is honored to be included on the Hooks Institute blog.

Soul City: Race, Equality, and the Lost Dream of an American Utopia by Thomas Healey

By Kat Robinson
Graduate Student, School of Public Health and Department of Anthropology, University of Memphis
Graduate Assistant, The Benjamin L. Hooks Institute for Social Change at the University of Memphis

This post is part of a series on the 2021 Hooks National Book Award Finalists.

Thomas Healey, Professor at Seton Hall Law School, brings us Soul City: Race, Equality, and the Lost Dream of an American Utopia, a 2021 Hooks National Book Award finalist.

Soul City chronicles the leaders of a local civil rights organization looking to establish a predominantly Black city in rural North Carolina during the 1970s. The goal was to move poor Blacks out of the South, in the hopes of fleeing poverty in search of prosperity in the North.

The core of Healey’s book demonstrates the “aspirations for self-determination, and economic autonomy, and equality among Civil Rights activists and leaders” from the late 1960s through the 1970s. Healey shows us how, after Martin Luther King, Jr.’s death, activists continued to strive towards justice and civil rights into the 1970s, fighting to ameliorate economic inequality, although, ultimately, their plan failed. In 1969, an emerging Soul City received financial support from the Nixon administration and North Carolina’s state government. Alas, even with the financial backing of both governing bodies, the project would prove unsuccessful some ten years into its existence. Healey delves into the various political and social obstacles that Soul City faced throughout its ten-year span and how these same obstacles exist today, impeding the Black population’s capacity for true economic autonomy and equality.

Healey remarks “[W]e’re still dealing with the issues of inequality today that we were in 1969…” He notes that unemployment rates are double the percentage of Blacks than Whites, similar to the 1960s. While the median wealth of Black families amounts to less than 15% of median white families (Bhutta, 2020); a stark reminder that the problems of the past are still apparent today.

Soul City revives a lesser-known story of the Civil Rights Movement; the journey towards economic autonomy and equality was not an uncommon goal, but Floyd McKissick strove for something more remarkable than imaginable. Healey details how a group of activists, led by McKissick, strove to end economic inequality and what obstacles they faced in their pursuits, and why they didn’t succeed, with the hope to develop a better understanding of the forces today standing in our way of economic equality.

 Healey refers to Dr. King’s book Where Do We Go from Here?, King’s last book before his assassination in 1968. In it, King asks readers where the Civil Rights Movement is to go next, hoping for economic prosperity and rights in education and housing. Floyd McKissick also believed the next step was full economic equality, and that Soul City was the solution to the inequities Black Americans faced. While many histories of the Civil Rights Movement end at King’s death, Healey argues that the end had yet come – King’s dream had yet to be realized. Although many recount King’s period as the golden years of the movement, Healey posits that McKissick’s plans cemented the call for economic equality.

Floyd McKissick was a famed Civil Rights leader and lawyer in the 1960s who headed the Congress of Racial Equality. He was born in Asheville, NC, and was the first Black student to attend the North Carolina Law School via a lawsuit brought by Thurgood Marshall. He participated in the first Freedom Rides of the 1950s in connection with the Journey for Reconciliation. After King’s assassination, he left his position as the leader of the Congress for Racial Equality to move home to NC to build Soul City, where he remained until his death in 1991.

About the Author

Kat Robinson began as a Hooks Institute Graduate Assistant in September 2022. Robinson is completing a dual-degree program in the School of Public Health and the Department of Anthropology. As a student at the University of Memphis, Robinson has participated in various student-led organizations, most notably Safety Net and the Diversity and Inclusion committee in the Anthropology Department. As a member of these organizations, Robinson learned the importance of organizing and working with community members to strive for equity and safety for students and Memphians.


Bhutta, N. A. (2020). Disparities in Wealth by Race and Ethnicity in the 2019 Survey of Consumer Finances. Washington: Board of Governors of the Federal Reserve System.

Until I Am Free: Fannie Lou Hamer’s Enduring Message to America – Hooks Blog Post From Author

By Keisha N. Blain

This blog post is part of our series on the 2021 Hooks National Book Award Finalists. For more visit

Fannie Lou Hamer’s story captures the contributions of a Black woman sharecropper with limited formal education and limited material resources—but an all-consuming passion for social justice. Born in Mississippi on October 6, 1917, Hamer was the youngest of twenty children. The granddaughter of enslaved people, Hamer worked as a sharecropper for much of her life. At the age of twelve, she concluded her studies at a local schoolhouse so she could help her family meet their growing financial pressures. Still, they remained trapped in poverty—the result of the exploitative nature of the sharecropping system and the violence used to maintain it. The difficulties of Hamer’s childhood extended well into adulthood. Despite her limited material resources and the various challenges she endured as a Black woman living in poverty in Mississippi, Hamer committed herself to make a difference in the lives of others.

Her life changed dramatically in 1962. On August 27, 1962, Hamer attended a local church service in Ruleville, Mississippi. At this meeting organized by activists in the Student Nonviolent Coordinating Committee (SNCC), Hamer learned of her constitutional right to vote as a citizen of the United States. She recognized that access to the ballot would help overturn unjust laws, replace corrupt elected officials, and shape local, state, and national politics. That night she began her political journey—relying on her radical honesty, boundless compassion, and unwavering resistance to racism and white supremacy.

Until I Am Free: Fannie Lou Hamer’s Enduring Message to America centers on Hamer’s ideas and political philosophies to demonstrate how they speak to our current moment. It posits that Hamer’s insights and political strategies during the 1960s and ’70s provide a blueprint for tackling a range of contemporary social issues. I had begun to write a book on Hamer in 2019 and thought I would finish it much later. However, the uprisings of 2020 as well as the global pandemic brought on a new sense of urgency for me to finish the book. During those difficult days, I found inspiration in Hamer’s words. The more I read Hamer’s words, the more clarity I found. Her vision for the world and her commitment to improving conditions for all people gave me a renewed sense of hope and purpose. I wanted to share that gift with others—and I firmly believe this is the opportune moment to share her story in a new way.

Until I Am Free is organized into six thematic chapters that explore Hamer’s perspective on faith, state-sanctioned violence, leadership, women’s rights, Black internationalism, poverty, and more. Together, these chapters grapple with one significant question: What might we learn, and how might our society change if we simply listened to Fannie Lou Hamer? One of her core messages—and the inspiration for the book’s title—that she delivered to audiences was, “Until I am free, you are not either.” Hamer recognized that no one could truly experience freedom if others were constrained. I use this framework as a starting point to remind readers that while the work of democracy is incomplete, the fight is certainly not over. As Hamer reiterated time and time again, we still have the power to make this inclusive vision a reality.

Walk with Me: A Biography of Fannie Lou Hamer – Hooks Book Award Finalist

By Kat Robinson, Hooks Institute Graduate Assistant

This post is part four of our series on the 2021 Hooks National Book Award finalists.

A cradle-to-grave biography, Walk with Me: A Biography of Fannie Lou Hamer explores the transformative life of Fannie Lou Hamer, from her early life picking cotton alongside her parents in Montgomery County, Mississippi through her later years as a voting rights activist until her death in 1977. Hamer used her experiences from sharecropping at 12 years old to revolutionize voting rights across the South; witnessing first-hand the discrimination and racism that Black people faced in the South, Hamer sought out like-minded individuals interested in justice and equity. With only a 6th-grade education, Hamer was a unique figure among her contemporaries like Dr. Martin L. King or Russell B. Sugarmon — and used her uniqueness to lead the voting rights movement into a direction it needed to go, against tremendous odds.

Historical and previously unavailable archives about Hamer were finally released, which Larson took advantage of when writing her biography. Hamer was the 20th child of Mississippi sharecroppers, born in 1917. She grew up in extreme poverty, experiencing malnutrition and limited access to healthcare and housing, and most horrendously she dealt with the impacts of white supremacy. These accumulated hardships led to 7 of her siblings dying before she was born. Out of this great loss was born a tight-knit family that Hamer would work to support throughout her youth.

Frustrated by lifelong crises, Hamer sought out a movement called the Student Non-Violent Coordinating Committee (SNCC) – an organization centered around nonviolent protests for civil rights. In 1961, the SNCC traveled to Mississippi to ask Black Mississippians what rights they felt they were missing as Americans – that answer was the right to vote. In the 1960s, only 5% of the Black Mississippian population was registered to vote because the violence faced when attempting to register was brutal. Hamer’s own experience of being fired for registering to vote further pursued her toward her mission. That mission was to travel throughout the South; illuminating the adversities and violence Blacks faced when attempting to register and leading voter registration initiatives. Not only in Mississippi but all over the United States.

Larson’s biography differs from others because it includes those recently released documents that have only resurfaced in the last 20 years, since the last biography about Hamer was released. Larson employs a feminist lens that places Hamer at the center of her life’s story: Hamer’s story not only differs from others due to being a woman in leadership within the civil rights movement but as a woman who had significantly less education than her male counterparts. Although Hamer was prevented from continuing her formal education, she possessed literacies and languages that her male counterparts did not. As a sharecropper and Black woman living in the South, Hamer experienced various gendered and racial harassment that informed her stance on equality and human rights. Unlike other activists and leaders, Hamer used her upbringing to inform her work.

Larson believes today’s reemerging threats on voting rights would deeply affect Hamer. Larson remarks on the current issues facing voting rights: “Today, her [Hamer] legacy stands strong. There are many lessons in her life that we can look to… to continue the fight for equality and justice here.”

This biography showcases the power behind “every day, ordinary people” who impact our nation’s society in positive ways. Hamer is a role model to show that the ‘average’ person can create great change. Larson says, “Across our country there are emerging leaders like Fannie Lou Hamer… They need support. That’s why her story is so important, because we need to be reminded that it happened before and we can do it again, and we can do better this time.”

Walk Don’t Sink: Journey through the Storms of Our Now

By Tony Coleman, associate pastor at First Congregational Church of Memphis

This blog was presented as a sermon at First Congregational Church of Memphis on June 26, 2022

As a child, I wasn’t afraid of many things. I loved the dark; Halloween was one of my favorite holidays; scary shows and movies for kids were some of my favorites. There was one thing that terrified me, though, one thing that would make my pulse jump and my face feel flush with anxiety: the Presidential Fitness Test. For those of you who may not be familiar with it, the Presidential Fitness Test was an annual event that took place in public schools across the country in which middle and high-school children would be forced to do pull-ups, sit-ups, push-ups, and more all under the eye of a gym teacher who would meticulously keep count of how many and how well the children did, in other words, torture. For a, as we graciously say in the South, big-bone-ded child like me, this whole measuring my fitness against what a “normal” child could do was terrible.

I stood in line and watched my classmates grabbed hold of the pull-up bar and easily do 4, 5, 10 reps. I watched them, seething with jealousy and dreading the moment that I would have to climb up and let everybody watch as I struggled to hang there and just barely bend my elbows, failing the test altogether.

Even worse than the pull-up bar, though, was the 10-minute mile test. Standing there on the starting line, next to kids whose little biceps were flexed and ready, having just completed 18 pull-ups with Terminator precision only minutes before, I could feel my heart beating, thumping with fear. The whistle blew. We ran, and I got lapped, again and again, as my classmates finished their 4 laps around the field before I was even halfway done with mine.

This went on for years.

Finally, in seventh grade, I decided that I would go up to Coach Williams, our gym teacher, to try and persuade him to let me sit out. To be honest, though, I was almost as nervous about talking to Coach Williams as I was about the Fitness Test. He was a tall man with thinning, grey hair who always wore a 90’s era track suit and a red whistle around his neck. He often seemed angry, for no reason, I thought. He would bark orders at us during warm-ups – “Get those hands up higher, Coleman, you call that a jumping jack?” Why yes, Coach Williams, yes I do, in fact – I wanted to say but didn’t.

When I told him that I wanted to skip the Presidential Fitness Test, that I was scared and embarrassed, that I’d be perfectly happy sanitizing basketballs or buffing out scuffs on the gym floor during that week, anything, he responded with, “You call that effort, Coleman? You think that’s good enough” Yes, yes, I do, Coach Williams – I wanted to say but didn’t.

“What you need to do, Coleman, is get angry. Get mad at that pull-up bar,” he told me. “Get aggressive. Get angry at that track out there – that’s how you’ll finish. I guarantee it.”

A part of me thought that maybe Coach Williams was on to something; maybe I did need to angry. Then, another part of me wondered what other kind of advice I could have gotten from a man I had once seen literally punch a basketball because it refused to go in the hoop. Getting angry was the way you get the job done, for Coach Williams. Feeling afraid? Feeling like you’ve failed? Get angry.

Of course, Coach Williams isn’t alone in holding that mindset, is he? Certainly, when it comes to fitness, we’re told that we can push our bodies further, we can tap into reservoirs of energy, we can exceed even our own expectations if we would only, “Be aggressive! Be, be aggressive!” Anger is the way to access unforeseen ability.

That’s true beyond the world of fitness and athletics, though, too. We see it in the way we describe so many tasks in our lives. We’re fighting COVID. We battle cancer. We fight for our rights. We battle our way through divorce. We fight our own desires for things we think we should not have. And so, in this way, the sum of our lives can be tallied up in all the little skirmishes we’ve won and lost.

Anger, then, is always in the background. In the fear we may feel when we get that diagnosis or start that journey towards weight loss or receive news of the latest devastating blow the Supreme Court has dealt, in the face of the fear we encounter, so often we’re told to get angry. Righteous outrage is the answer. Channel your fear into anger and your anger into energy so that you can keep on, push through, and claim victory in the end. Don’t give up; get angry!

It’s in that spirit that we might be tempted to read what happened with Peter and Jesus today. Jesus is there immersed in a miracle, standing on water. Peter sees this, is shocked and frightened, and says, if that’s really you, Jesus, tell me to come out there and join you. Jesus does, then Peter does. Peter actually walks on water. He sets aside his fear about the storm and his fear about seeing a man walking atop crashing waves, and he takes a step into a miracle.

Before long, though, he gets distracted by those crashing waves. He gets frightened all over again by the storm raging around him, and he begins to sink. Flailing around in the waves breaking beside, behind, and in front of him, he reaches out for Jesus’s hand, and Jesus catches him. “You of little faith,” Jesus then says to him, “why did you doubt?”

Now, at first glance, I’m inclined to hear Jesus saying a version of what an angry gym teacher might say to his chubby student: “You call that effort, Peter? You need to push your faith. You need to try harder. You need to get angry at doubt.” Maybe.

But, if we take a step back, if we really look at what’s happening here, if we let ourselves get absorbed into the miracle taking place, we see that this isn’t just about pushing through. This is about seeing differently. Jesus was inviting Peter into something bigger than sheer force of will, something bigger than even just walking on water. Jesus was inviting Peter into a whole other vision of the world, a world that shatters expectations and explodes logic, a world that is governed by something bigger than scary storms and angry effort, a world in which we reach our goals not by force alone but with faith in God.

Peter started sinking, then, because even though Jesus was showing him what the world could be, all Peter could focus on was what he thought the world was. Jesus was calling Peter into a reality that defied gravity and shamed his shallow expectations, but, before too long, Peter got lost in his own sense of what was possible, in his own sense that all he could do, when it came to raging storms, was sink.

Jesus was not saying face your fears in the midst of that storm and get angry. Jesus was saying, in the face of your fears, you can step into my vision for what’s possible, and you can discover in that miraculous potential.

Friends, there are some journeys whose ends we will never reach if all we have fueling us is anger. To be sure, anger has its uses. Anger can give us energy. Anger can make our voices strong when they may otherwise be weak. Anger can lead us to take action when we might otherwise be still. Anger can help us sprint. But, anger can’t get us through 6 months of chemo. Anger can’t help us figure out what’s best for our kids when a relationship falls apart. Anger can’t make a 200 pound 12-year-old run a mile in under 10 minutes. Anger can’t sustain a movement. Anger can’t help us travel from the boat into the miracle. Anger can help us sprint, but it can’t help us run a marathon.

“The arc of the moral universe is long, but it bends toward justice.” Chances are, if you run in any kind of progressive circles, you’ve heard or read this quote. It’s on t-shirts. It’s said in political speeches. It serves as the little taglines for emails. It’s tweeted and instagramed and Facebooked. It’s almost everywhere nowadays, and, because of that, we may lose sense of what these words were actually doing when Dr. Martin Luther King, Jr. said them.

You see, these were not words for a short run. This was not the sentiment of a person just trying to push through. These were the words of a man and a movement who knew the reality of setbacks and the seemingly impossible task before them. It’s a set of words that reach beyond policy that dig underneath ideology that take hold of a truth that is bigger than the latest failure.

MLK’s words said that while the forces of white supremacy and oppression may win this battle and the next one or even the one after that; while I may even lose my life in pursuit of the goal of justice, at the end of the day, when it’s all said and done, I know that we are going with the flow of the universe, rather than against it. These words show that the people in this movement carried a confidence that the very shape of existence, the very fabric of reality, the plain ol truth of the way things are, was caught up in the very same thing that they were – justice, healing, and love. Their vision of the world was one that defied expectations and lent them a confidence that achieving their goal was, in a very real way, inevitable.

Dr. King was reiterating the same core point that Jesus was sharing with Peter – in order to make the distance, in order to walk rather than sink into fear or anger, in order to do the work of living with intention and dignity in the face of the incredible storms that rage all around us, in order to do that, we have to see that the world is so much bigger than fear might lead us to think. We have to see that while battles can be won or lost, faith operates according to an entirely different logic.

With faith that the world bends toward justice, with faith that God will reach out and grab us when we’re starting to sink, with faith that we can stand on top of the waves rather than get swallowed up in them, that is the only thing that will take us through. That kind of faith is bigger than a person. It’s bigger than a court. It’s bigger than a treatment plan or a solution, bigger than a politician or an idol. That kind of faith is about stepping into a world that loves us, that holds us, that carries us in our walk, no matter what stumbling blocks we may encounter along the way. That kind of faith says to a chubby pre-teen: there is a vision of the world that loves you even if it takes you 22 minutes to run a mile rather than just 10; there is a vision of the world in which sickness and death are not things you battle but realities you learn how to hold. There is a vision of the world, in which you don’t have to get angry and try to do it all on your own.

We’ve got a long journey ahead of us, friends. We’ve got problems, personal and global. Problems with our country, problems with our bodies, problems with our building. Problems that are so big and so deep, we may not know what to do or even where to begin. We may feel scared and yes, we may feel rage. But, the truth is, neither fear nor anger are going to sustain us for this journey. This journey’s too long; the road is too hilly; the storms are too fierce. What’s going to help us keep going, what’s going to help us get out of the boat and step into uncertainty, what’s going to help us travel the distance of the universe’s long moral arc, what’s going to help us in our fear and our grief and our anxiety, what’s going to help us make it—is a vision. What’s going to help us is a vision of the world, rooted in faith and repeated throughout history, that has enabled human beings to do incredible things. It’s a vision that defies expectations and makes room for miracles. It’s a vision that focuses on the path rather than the storm. It’s a vision that will help us to walk rather than sink.

Let us be together, friends, and let us envision with hope, courage, and faith.


Tony Coleman is an associate pastor at First Congregational Church of Memphis (aka First Congo). He grew up in a bi-racial, working-class family that called several Memphis neighborhoods home. After high school, however, he moved to the Northeast for college and graduate school. Tony is passionate about writing, art, and creativity – and how all of these can facilitate and sustain social change. He has been published in the Christian Century and various online magazines including Conversation X. In addition to matters of faith and the spirit, Tony writes and thinks about food, particularly baking and eating it.

The Young Crusaders: The Untold Story Of The Children And Teenagers Who Galvanized The Civil Rights Movement

By B.P. Franklin

Book published by Beacon Press, Boston, MA, Winter 2021

This post is part of our series on the 2021 Hooks National Book Award Finalists.

Some of the most iconic images of the Civil Rights Movement are those of young people engaged in social activism: The Little Rock Nine being escorted into Central High School in 1957 by soldiers—or children and teenagers being attacked in 1963 by police in Birmingham with dogs and water hoses. While the contributions of Dr. Martin Luther King Jr., the NAACP, and the Student Nonviolent Coordinating Committee have been well chronicled, the crucial roles of children and teens are now placed at the forefront in The Young Crusaders.

V.P. Franklin delivers a thorough reexamination of the efforts of children and teenagers to challenge legal segregation, employment discrimination, educational inequality, and racialized violence beginning in the 1930s. His groundbreaking narratives draw on examples in nearly fifty cities and fifteen states, from Alabama to Wisconsin. Franklin details the student activism behind the successful civil rights campaigns that brought about the end of Jim Crow practices throughout the nation.

For instance, the largest civil rights demonstration in US history was not the famous “March on Washington” in August 1963, but the system-wide school boycott in New York City on February 3, 1964, when over 360,000 elementary and secondary school students went on strike and thousands attended Freedom Schools.  Multiracial community groups organized the school boycotts in New York, Chicago, Cleveland, and other cities, and African American, Puerto Rican, and white students participated, calling for the desegregation of public and private education and public accommodations. In addition, children and teenagers mounted their own walkouts, marches, and civil rights demonstrations demanding “Freedom Now.”

This account of the courageous actions of these unheralded young people fundamentally transforms how we understand the Civil Rights and Black Power Movements and serves as a playbook for youth-oriented protest movements such as The Climate Strike, March For Our Lives, and Black Lives Matter, reigniting in the twenty-first century the next wave of social and political activism.

About the Author
V. P. Franklin is Distinguished Professor Emeritus of History and Education at the University of California, Riverside. Between 2001 and 2018 he served as the Editor of the award-winning Journal of African American History and is the author or co-editor of many books including Living Our Stories, Telling Our Truths: Autobiography and the Making of the African American Intellectual Tradition, and Sisters in the Struggle: African American Women in the Civil Rights-Black Power Movement.

The Citizenship Education Program and Black Women’s Political Culture | Hooks National Book Award Finalist

This post is part of our series on the 2021 Hooks National Book Award finalists, written by the authors.

By Deanna M. Gillespie

The Citizenship Education Program and Black Women’s Political Culture retells the story of the civil rights movement to foreground southern Black women’s formal and informal networks.  From 1957 to 1972, the CEP empowered thousands of Black women to teach their friends and neighbors to read and write.  This was subversive work in the Jim Crow South.  Literacy skills shifted the balance of power because literate Black people gathered information, registered formal complaints, demanded equal treatment, signed their names, and they stood in front of white county clerks and took the literacy test required for voter registration.  Across the South, Black women organized CEP classes in church halls, community centers, living rooms, and beauty salons. These makeshift classrooms were incubators for collective action.  Gathering twice a week for three months, local people organized to address long-standing inequities in social welfare and community well-being, affordable and accessible health care, equality and justice, education, and criminal justice reform.

Hear from Author Dee Gillespie on Our YouTube

Black women taught CEP classes because it was familiar work.  It was “women’s work.”  In segregated all-Black schools designed to teach racial inferiority, Black teachers told and retold stories of resistance and achievement.  In local communities, mothers, grandmothers, aunts, and neighbors guided children safely through Jim Crow’s unforgiving rules.  They taught the Bible on Sunday mornings and carried the message through the week.  Theirs was – and is – a political culture built on family and community, constant struggle, and the unshakable belief in human dignity and a just and better day.

Organized into four case studies, the book traces the CEP’s evolution from its beginnings on the South Carolina Sea Islands, to expansion into Savannah and southeastern Georgia, out to Mississippi, and back to Selma and Alabama’s Black Belt.  In the final chapter, I examine the program’s decline after 1965, a reminder of how quickly pieces can scatter when political winds shift.  By the time the CEP ended in 1972, the program had trained over two thousand local leaders who organized 7280 classes, reaching over 26,000 people. As teachers and students fanned out, they influenced an estimated 95,000 people to register and to vote.

The CEP tapped into a deeply-rooted gendered political culture and recent elections provide evidence of its enduring nature.  In Fall 2020, the United States held a presidential election during a global pandemic.  Across the country, voter participation soared through expanded use of mail-in ballots, drop-boxes, and early voting periods.  Grass-roots organizing and mobilization turned reliably-“red” Georgia “blue” for the first time in a generation.  News coverage shined a bright spotlight on Black women’s clubs, organizations, and local leadership.  Likewise, inDecember 2017, headlines announced Democrat Doug Jones’s victory for Alabama’s contested U.S. Senate seat.  Within hours, #BlackWomen trended on social media.  In post-election interviews, these women cited concerns about the future of social welfare programs, health care, education, and criminal justice reform as motivation for political action – the same issues that enlivened CEP classes a generation before.

Flatez Dyson: Success through HAAMI

By Anjali Gahlaut

Headshot Flatez Dyson
Flatez Dyson

The Hooks Institute’s “Where Are They Now?” series explores how the institute has shaped the lives and careers of former alumni. The series uncovers valuable experiences, skills, and lessons that former alumni acquired during their time at Hooks.

Flatez Dyson is an alumnus of the Hooks African American Male Initiative (HAAMI) at the Benjamin L. Hooks Institute for Social Change. HAAMI seeks to enrich the collegiate experience for its members as well as improve graduation rates for African American male undergraduate students at the University of Memphis. Dyson first got involved with HAAMI in 2015 and was a member of the program’s first cohort. He states, “I thought it was a great opportunity to get involved with a program that was greater than myself and be involved with other like-minded individuals.”

During his time with HAAMI, Dyson attended multiple special events where he was able to speak and network with people of various professions. Public speaking is a skill Dyson has integrated into his professional life as a supply chain analyst at Monogram Foods. “Those experiences through HAAMI built up my confidence level in public speaking to where I could speak to individuals at different levels within an organization. At my current job, I report directly to the Chief of Supply Chain, so I have daily conversations with upper management. It [HAAMI] has definitely benefited me and my career in being able to speak boldly and proudly about the things I stand by.”

Hear from Flatez Dyson

In 2016, the Hooks Institute’s annual “Join Hands for Change” gala honored the accomplishments of University of Memphis alumnus Marvin Ellison, who was CEO of JCPenny at the time. As a member of HAAMI, Dyson was able to speak at the gala and have one-on-one conversations with Ellison. Ellison was impressed by HAAMI and invited its members to Dallas, Texas, where they were able to participate in workshops and tour JCPenny’s headquarters. Dyson states, “For me, that trip had a lot of firsts in my career: it was my first time going to Dallas, it was my first time being on a plane, and it was the first time I met someone in a professional role of his caliber that looked like me. That was a very inspirational moment for me.”

HAAMI supports its students towards graduation through academic achievement, career development, and personal development. “HAAMI is a great organization to be a part of if you’re looking to be a part of an organization that really believes in its students, that puts its students first, and really pushes its students to thrive. I feel like it’s very family oriented. The staff and faculty members give you a nudge whenever you need that nudge, but they also give you career advice and parental advice and teach you life lessons. It’s a wonderful program to be a part of.”

About the Author

Anjali Claire Gahlaut graduated from White Station High School in Memphis, Tennessee and is now a rising sophomore and prospective History major at Vassar College in Poughkeepsie, New York. Anjali spent the summer of 2022 as an intern at Hooks and joined the team to gain a deeper understanding of Memphis’ history and its impact on marginalized groups.