The Federal Trade Commission could soon begin investigating big tech companies whose algorithms result in racially biased ads or search results, signaling the Biden administration's willingness to directly challenge social media companies' main sources of revenue.
Companies like Facebook use algorithms that target ads to narrow sets of consumers by processing huge quantities of data, including user activity, location, and facial recognition. Facebook has faced scrutiny for its micro-targeting tools, which let landlords market their apartments exclusively to white people and let employers target certain demographics for job postings.
The FTC's statement, released on the agency's official blog, sets the stage for the first high-profile battle between the Biden administration and big tech. Big tech companies draw much of their revenue from advertising targeted to users' personal information. That function could be sharply curtailed if the FTC considers ad targeting discriminatory.
The post cites laws that prohibit companies from using "racially biased algorithms." The commission bans "unfair or deceptive business practices," while other laws bar discrimination in credit or housing based on race or sex. The blog post's author, FTC staff attorney Elisa Jillson, warns companies that if they fail to hold themselves accountable, "the FTC may do it for you."
The FTC is one of several federal agencies that have begun investigating big tech algorithms. In December, the agency forced nine major social media companies to disclose how they target content to consumers. In 2019, the Department of Housing and Urban Development charged Facebook with discrimination for offering tools to help target ads that excluded "women in the workforce, moms of grade school kids, foreigners, Puerto Rico Islanders," and others.
Facebook in particular has received criticism for its opaque algorithms. Researchers at the University of Southern California found this week that the site disproportionately filters job postings based on user gender and "learns and perpetuates" existing bias.
In February, a judge approved a $650 million settlement between Facebook and users whose faces it scanned and stored without permission. And the Supreme Court is refusing to block a suit that claims Facebook's trackers on third-party websites are illegal. Those trackers are a key tool for Facebook, allowing the site to integrate information about users browsing the web into its own algorithms.
But the FTC may have trouble regulating big tech companies. "You don't announce something like this in a blog post,'’ American Enterprise Institute privacy scholar Jim Harper told the Washington Free Beacon. "If they're saying what it looks like they're saying, it's a far bigger deal that should be carefully thought through."
It's unclear how the FTC will put some of the language in its blog post, including the claim that it can challenge algorithms that "do more harm than good," into operation. The FTC does not have clearly defined "basic standards, benchmarks, and measurement tools" for evaluating algorithms for bias, noted artificial intelligence researcher Jon Stokes, and neither does the broader artificial intelligence industry.
FTC commissioner Rohit Chopra has expressed concern that the commission isn't ready or willing to take on major companies. In written testimony submitted for a congressional hearing Tuesday, Chopra warned that "the FTC is unwilling to take meaningful accountability measures" when companies like Google and Facebook repeatedly violate the law.
Reached for comment, an FTC spokeswoman told the Free Beacon that "combating algorithmic discrimination" is a priority for acting FTC chairwoman Rebecca Kelly Slaughter.