Dems introduce bill targeting bias in algorithms

Congressional Democrats on Wednesday introduced legislation that would require companies to correct algorithms that result in biased or discriminatory actions that harm Americans.

The measure, sponsored by Sen. Ron Wyden (D-Ore.) and Rep. Yvette Clarke (D-N.Y.), would authorize the Federal Trade Commission (FTC) to enact regulations requiring companies under its jurisdiction to assess the impacts of sensitive automated decisions for their effects on bias, discrimination, privacy and accuracy, according to Wyden’s office.

{mosads}Sen. Cory Booker (D-N.J.), a 2020 presidential candidate and co-sponsor of the legislation, said in a statement Wednesday that discrimination from decades ago can be “significantly harder to detect in 2019: houses that you never know are for sale, job opportunities that never present themselves, and financing that you never become aware of—all due to biased algorithms.”

Several tech privacy and civil rights groups endorsed the measure, including Data for Black Lives, an activist group that combats uses of data science it says contribute to racial discrimination.

“We know first-hand the harmful impact that automated decision systems have on parents fighting for access to quality education, black mothers engaging health systems in how to provide care that protects their newborns, and activists fighting against community disinvestment and deprivation,” Data for Black Lives said in a statement. “Beyond regulation, we are hopeful that this legislation will lead to a broader discussion about the tremendous potential for data systems, if used ethically, to uplift, empower, and democratize our communities.”

The bill would apply only to firms already subject to FTC regulations with annual revenue that exceeds $50 million. However, it would apply to all data brokers or companies with data on more than 1 million consumers or devices, regardless of revenue.