Bill would make firms disclose AI decision-making risks

What does this do?

This bill would require certain large companies to study and document how their AI and automated decision-making systems affect people before and after using them. Companies would have to test for bias across groups defined by race, sex, age, and similar factors, consult outside experts and affected communities, and try to fix serious problems they find. The FTC would keep a public database of company reports so consumers and researchers can see which algorithms are in use and how people can challenge decisions made about them.

Who does it affect?

The rules would apply to large companies above certain revenue or data thresholds that use algorithms to make or influence important decisions about loans, jobs, housing, healthcare, or education. The FTC would enforce the law, state attorneys general could also sue companies that break it, and a new Bureau of Technology would be created inside the FTC to help carry it out.

Why does it matter?

Without these requirements, automated systems that affect major life decisions could go unexamined for bias or harm. This bill creates a process meant to make those systems more visible and to give people a way to challenge decisions made about them.

Where does it stand?