Skip to content
Search

Latest Stories

Top Stories

Why are Black women becoming the hidden figures in AI?

Katherine Johnson receives the Presidential Medal of Freedom from President Barack Obama

President Barack Obama presents Katherine Johnson with the Presidential Medal of Freedom in 2015. She would not become widely recognized until the release of "Hidden Figures" in 2016, decades after she made groundbreaking contributions to the space program.


Kris Connor/WireImage

Turner Lee is a senior fellow and the director of the Center for Technology Innovation at the Brookings Institution. She has created a new AI Equity Lab at Brookings to interrogate civil and human rights compliance within emerging models.

The White House recently issued an executive order to ensure the safety, security and equity of artificial intelligence technologies in the United States and abroad. This call to action was reaffirmed when Vice President Kamala Harris spoke at the AI Safety Summit in the United Kingdom and she emphasized the need for expanded oversight, and possible AI regulation over machines and models that result in adverse consequences.

For decades, an increasing number of experts, including other Black women in AI, have generated research and policy recommendations around necessary civil and human rights protections, improved data quality, and the lawful compliance of AI and other emerging technologies. While the bold prominence of the vice president at the summit set the stage for more diverse voices, Black women are largely being excluded from these and other high-profile dialogues, which disadvantages any conversations on the design and deployment of more equitable AI.


Black women have led in AI policymaking before its growing popularity and interest. In 2019, Rep. Yvette Clarke (D-N.Y.) issued the first version of the Algorithmic Accountability Act with her colleagues to interrogate automated decisions driving housing, creditworthiness and hiring outcomes. In the same year, and after her former Speaker Nancy Pelosi was the subject of altered videos, Clarke also introduced the first version of her DEEPFAKES Accountability Act, which would go after video creators who falsified video content and contributed to disinformation, especially around the time of the 2020 election.

Sign up for The Fulcrum newsletter

After presenting revised updates to her proposed legislation every year following, Clarke’s 2023 versions of both are up again for House consideration. Other Black women in policy like former White House official Alondra Nelson led the development of the first ever national Blueprint for an AI Bill of Rights in 2022 that pushed the country closer to more equitable AI.

The indelible marks made by these women have not gone unnoticed. But they are joined by an increasing number of other Black women in AI, whose work includes formidable research and pragmatic policy proposals that bring lived experiences to AI design, deployment and oversight. Despite impressive backgrounds as computer and data scientists, criminologists, sociologists, curators, policy professionals, tech entrepreneurs, and social activists, many of these women are not prominently appearing in highly visible roles and conversations or being asked by policymakers to bring their findings to congressional hearings focused on AI regulation, technical cadences, or harms to consumers and democracy.

Between March 8 and Nov. 29, Congress held various hearings on AI that covered topics relating the workforce to AI’s role in modern communications. Black women were sparsely presented as witnesses to testify on the subject. Of the 125 witnesses who participated in the 32 hearings, only two were Black women, one of whom was me. Until recently, Senate Majority Leader Chuck Schumer has worked to increase the number of Black women participating in his recurring AI Insight Forums, which will undeniably influence the direction of national AI self-regulatory policies, and more prescriptive regulation.

The reality is that when Black women’s voices are amplified, it is usually to discredit their contributions and achievements. AI pioneer Timnit Gebru, a former Google researcher, was terminated after announcing the company was silencing marginalized voices. Algorithmic justice advocate and bestselling author Joy Buolamwini exposed the racial inequities embedded in the design and use of facial recognition technology by government agencies. Yet, law enforcement continues to use it with little regard for the false arrests of Black people due to its technical misidentifications around skin color.

When Black women are not being attacked for their similarly poised expertise, they are relegated to the status of being hidden figures in science and technology – where they are more of an afterthought. In 2016, the world came to know the late Katherine Johnson, a retired NASA scientist and mathematician in orbital mechanics, through a bestselling book and movie both titled “Hidden Figures.” At the age of 96, Johnson’s expertise was finally recognized, especially her work on many critical space missions in her decades-long career, including the aversion of the near-fatal collision of the Apollo mission. Four years after we learned about her significant role in the country’s space programs, she died at the age of 102.

Black women in AI should not be attacked or become hidden figures where their ideas and concepts are invoked by others in privileged conversations and rooms.

If future AI is going to be responsibly designed and deployed, Black women must be included to stop and control how these technologies further racialize or criminalize certain communities. Their participation can motivate more inclusive design and products in AI that better serve and involve the public. As more corporations are dismantling diversity, equity and inclusion programs that will erode Black women’s visibility in board rooms, and other prominent leadership roles, this call for participation and input are both critical and timely.

Right now, it feels like Black women, whose thought leadership comes packaged in so many disciplines and work streams are largely invisible despite their established expertise. If the federal government, and the companies at the forefront of AI development, desire to achieve the goals of safety, equity and fairness in AI, then it must see and include Black women on Capitol Hill, in university research labs, in civil society organizations and think tanks, as well as in industry war rooms and boardrooms. We must not be memorialized later as hidden figures for the decisions that ultimately got made today about the governance, development and deployment of more ubiquitous AI.

Read More

Main Street AI: AI for the People

An illustration of AI chat boxes.

Getty Images, Andriy Onufriyenko

Main Street AI: AI for the People

When Vice President J.D. Vance addressed the Paris AI Summit, he unknowingly made a strong case for public artificial intelligence (AI) infrastructure. His vision—of AI that empowers workers rather than displaces them, enables small businesses to compete with tech giants on a level playing field and delivers benefits to all Americans—cannot be achieved through private industry alone. What's needed is nothing less than an AI equivalent of the interstate highway system: a nationwide network of computational resources, shared data, and technical expertise that democratizes access to this transformative technology.

The challenge is clear. The National AI Opinion Monitor reveals a stark digital divide in AI adoption: higher-income urban professionals increasingly leverage AI tools to enhance their productivity, while rural and lower-income Americans remain largely locked out of the AI economy. Without intervention, AI threatens to become another force multiplier for existing inequalities.

Keep ReadingShow less
Data-based checks and bicameral balancing of Executive Orders
shallow focus photography of computer codes

Data-based checks and bicameral balancing of Executive Orders

The flurry of Presidential Executive Orders attracted plenty of data-based checks in the media. The bad propaganda, rollbacks, and a dip in the President’s approval rating may have been avoided if the US Constitution mandated the Whitehouse to do similar checks before initiating the Executive Orders.

Mandating data-based checks on executive orders ensures that decisions made by the President are rooted in evidence and have a clear, justifiable basis. Data-based checks would ensure that executive orders are issued only after they are scrutinized on their merits, impact, and alignment with the public interest. These checks help prevent orders from being issued on personally or politically motivated priorities or unsubstantiated claims.

Keep ReadingShow less
TikTok: The Aftermath
File:TikTok app.jpg - Wikimedia Commons

TikTok: The Aftermath

When Congress passed PAFACA (Protecting Americans from Foreign Adversary Controlled Applications), they should have considered the consequences. They apparently didn’t.

With approximately 170 million users, what did politicians think would happen when TikTok actually went dark? Did Congress consider the aftermath? President Trump is trying hard to find a way to keep TikTok from going dark permanently, but he likely won’t succeed.

Keep ReadingShow less
Amid Trump’s War on LGBTQ+ Teens, Social Media Platforms Must Step Up
rainbow drawing
Photo by Alex Jackman on Unsplash

Amid Trump’s War on LGBTQ+ Teens, Social Media Platforms Must Step Up

With Trump’s war on inclusion, life has suddenly become even more dangerous for LGBTQ youth. The CDC has removed health information for LGBTQ+ people from its website—including information about creating safe, supportive spaces. Meanwhile, Trump’s executive order, couched in hateful and inaccurate language, has stopped gender-affirming care.

Sadly, Meta’s decision in January to end fact-checking threatens to make social media even less safe for vulnerable teens. To stop the spread of misinformation, Meta and other social media platforms must commit to protecting young users.

Keep ReadingShow less