Artificial Intelligence

Washington DC Attorney General launches algorithmic bias bill

Law would levy $10,000 per infection for algorithmic bias
Image: Shutterstock via Dennis

13 December 2021

Washington DC Attorney General Karl A Racine has proposed legislation that would prevent algorithmic discrimination. The Stop Discrimination by Algorithms Act (SDAA) would prevent companies from making automated decisions that discriminate against marginalised groups in the District.

The Center of Privacy and Technology and Communications and Technology Law Clinic (both part of Georgetown University Law Center) helped draft the bill, in conjunction with civil rights nonprofit Color of Change.

The Act forbids discrimination by algorithms that make decisions about credit, education, housing, employment, and public services. Companies must also inform individuals about how they use algorithms to make those decisions.




Organisations must document how they build their algorithms and how they make decisions. They must file annual audit reports with the Attorney General in which they check their algorithms for discrimination.

The proposed legislation comes with teeth, as the it allows the Attorney General to penalise companies that violate the rules by up to $10,000 per violation. It also permits private lawsuits.

The text of the bill warned that algorithmic decision-making systems that don’t account for bias can harm marginalised groups. “Despite their prevalence and the potential problems they pose, algorithms are poorly understood by most individuals, in part because of the many entities involved and the lack of accountability among those entities,” it said.

The move prompted support from organisations and individuals including Timnit Gebru, the AI researcher who complained that Google silenced her concerns about its AI algorithms.

“At this point it should be clear that multinational corporations will not self regulate,” said the scientist, who this month launched the Distributed Artificial Intelligence Research Institute. “To the contrary, they push out people with the slightest criticism of their proliferation of harmful systems. Without laws requiring companies to assess the potential for discriminatory impact of their algorithms, what they do instead is eject people like me who attempt to do that internally, even though this was literally in my job description.”

The proposed legislation mirrors efforts at the federal level. In 2019, Democrats introduced an Algorithmic Accountability Act that would have pressed the Federal Trade Commission to create rules for assessing bias in automated systems. In May this year, Democrats tried again with the Algorithmic Justice and Online Platform Transparency Act of 2021.

The FTC has already warned organisations to use AI responsibly, threatening potential action for those that fail to do so. It has issued guidance for companies to follow.

© Dennis Publishing

Read More:

Comments are closed.

Back to Top ↑