The Interconnectedness of Artificial Intelligence, Black People and Stolen Data

Photo by Osarugue Igbinoba on Unsplash

For over 400 years, the bodies, lives and data of Black Americans have been treated as raw material. We have been extracted, studied, and categorized in service of systems that were never designed with Black wellbeing in mind. Race-based medicine and data compilation have long been instruments of racial injustice, their unethical practices leaving indelible marks on African American communities across generations (Cerdeña et al., 2020).

In the 21st century, it is evident that Artificial Intelligence is no longer a futuristic possibility. It is already here, knocking on every door, and quietly shaping how we work, connect and consume information. Nearly every major technological platform harvests user data to train its models, and yet, the question that demands to be asked is one that few are willing to sit with: is today’s data infrastructure truly different from unethical research practices that have historically dehumanized Black lives — or it is simply wearing a new face?

This editorial seeks to reframe that question. By tracing the throughline between America’s past of anti-Black scientific exploitations and the data practices embedded in modern AI systems, it aims to bring forward what is too often obscured — that the digital age has not broken from this history, but inherited it. Awareness is the first act of resistance. This is an invitation to see it clearly .

The compelling story of Henrietta Lacks is interconnected to the data footprint of AI primarily as a cautionary ethical archetype representing the exploitation of personal, biological data without consent. Henrietta Lacks sought care at the John Hopkins Hospital in the 1950s for what would later be discovered as cervical cancer (Skloot, 2010). HeLa cell lines continue to thrive to date, helping with the advancement of science and medicine. The cell line has been used to make discoveries for cancer, HIV, Ebola and tuberculosis in‐vitro fertilization (Sodeke & Powell, 2019). More recently, HeLa cells have been used for human genome studies, virology and for the development of the COVID‐19 vaccine (Johns Hopkins Medicine, 2022). The story of Henrietta Lacks was not an isolated incident nor is the focus only on the flawed practices of cell retrieval, biobanking, and lack of informed consent, but the reciprocal pattern of exploitation, surveillance and capital gain from the lives of Black people, dating back to the early 1800s to present day. Biological data has been extracted without consent for years, and has generated billions for pharmaceutical companies and other science organizations. Yet, today’s behavioral and personal data extracted without meaningful consent generates the same capital that continues to advance AI development.

This pattern did not emerge from medicine alone, it was preceded by centuries of legally sanctioned surveillance and control of Black bodies, rooted in a system that has never required the consent of Black people to operate. The Lantern Laws of 1712 forced surveillance in the accordance of the state intended to control crime prevention, but surveillance has never been about crime prevention for Black communities, it was only framed that way. Instead it was another perceived entitlement to monitor, access, and control Black life without consent. These so-called “laws”presumed Black guilt and demanded Black visibility. One might assume the opportunity to surveil any human being would end during the civil rights era, it was only modernized. But,what carried forward from 1712 was not just the practice of surveillance, it was the foundational logic underneath it. That logic found a new instrument. The lantern became a lens. The lens became an algorithm. The algorithm, trained on incomplete and unrepresentative data, did what the Lantern Laws were designed to do — look at a Black face and render a verdict before a word was even spoken.

AI models are built on datasets that structurally underrepresent Black life, encoding absence as a form of bias before any algorithm is even written. Though the real-world consequences of unregulated facial recognition technology on Black Americans within the criminal legal system has been documented, the urgency for change continues to be deferred. Porcha Woodruff was eight months pregnant when she was arrested for carjacking after Detroit police ran surveillance footage through facial recognition software, returning a photo as a misidentification — then questioned for eleven hours, before the case was dismissed due to insufficient evidence. Woodruff was not alone. To date, every documented case of a wrongful arrest from facial recognition technology, involved a Black individual.

These are not coincidences, but the predictable output of algorithms trained on data that never was meant to adequately represent Black faces to begin with. The real question is, how much longer will the algorithm, data practices and the absence of any federal policy regulation continue to fail us? What is needed is not simply awareness of the harm, but accountability for how it is built. Who is controlling the data, how it is collected and whose consent was never asked for in the first place.

Regrettably, society is shifting toward passive consumption. We are prioritizing results while neglecting to question the methods behind them. If one has not realized it yet, artificial intelligence does not simply learn. At the foundational core of every agentic model lies a complex data pipeline, aggregating your clicks, posts, searches, your input, all in the name of enhancing your experience. What is rarely disclosed, is that you are not just the user, you are the resource. In the great words of the late Notorious B.I.G, “If ya’ don’t know, now ya’ know”. Your behavioral and personal data fuels the innovation pipeline of every AI development in the same structural way. Though your direct input is not sold, the agentic models trained on it continuously power features that drive revenue for the world’s largest technology companies. The mechanism has changed. The extraction has not.

Are we doomed? Well, not entirely. What this moment in time demands is structural accountability, not just awareness of the harm in which it causes. Data provenance is not a technical luxury, it is the foundational requirement for accountability. In every case documented in this piece, the absence of traceable data makes accountability impossible. Lacks’ family has spent years unable to challenge the use of her cells because there was no record of her biological data. The Detroit Police Department deployed a facial recognition system that misidentified Woodruff, but no record existed to explain which dataset produced the match. Data provenance provides the paper trail.

Medical standards, after years of exploitation, have adopted informed and documented consent as a non-negotiable ethical requirement. However, that same standard does not currently exist in AI systems today, at least from any meaningful or enforceable form for the behavioral and personal data being harvested. Informed consent architecture applies both ethically and digitally, requiring that users are informed on what data is being collected, how it is being used, which models it will train and the retain the right to withdraw that consent at any point. Historically, black communities data has been extracted without knowledge, without compensation, and without recourse. The same protection that should have been given to Henrietta Lacks in 951 is owed to every person whose digital life is being aggregated, monetized, and deployed against them today.

The through line of this piece is not subtle. The architecture of extraction has remained intact across centuries, updating its instruments while preserving its logic. Black bodies have been surveilled, studied, and monetized by systems that have never required their consent to operate. In this chaotic age of artificial intelligence, that architecture has found its most scalable form yet. Data compiled without meaningful consent, models trained on datasets that structurally exclude the communities they are deployed against, an unregulated environment that has yet to demand accountability. Put simply, this is not progress wearing a neutral face. It is the same system, encoded. Awareness is the first act of resistance, but it cannot be the last. To see this clearly is to accept that the question is no longer whether these systems cause harm. The evidence is documented, named, and on record. The question is whether we will continue to look away.