Trump officials challenge tech over housing discrimination – By Emily Birnbaum (The Hill) / April 2 2019
The Trump administration is stepping up its scrutiny into whether Silicon Valley giants are enabling housing discrimination online.
The administration has already brought charges against Facebook and is investigating both Twitter and Google.
The moves mark a significant escalation in regulators’ efforts to apply federal civil rights law to online advertising and is creating a surprising alliance between watchdog groups and Housing and Urban Development Secretary Ben Carson. Experts say that if regulators step up enforcement, companies could be forced to significantly overhaul their lucrative ad-targeting strategies.
The Department of Housing and Urban Development (HUD) surprised tech watchers last week by charging Facebook, one of the world’s largest players in online advertising, with encouraging and enabling housing discrimination through its targeted advertising practices. The agency argued that the practices allowed realtors and landlords to exclude certain demographics from seeing housing ads.
HUD also accused Facebook of excluding women, minorities, disabled populations and other protected groups from seeing certain ads even when the advertiser did not intend to do so.
A HUD official confirmed to The Hill that the agency at the end of last year sent letters requesting meetings with Google and Twitter to talk about their housing advertisement practices.
Lawmakers hailed the move and said they would be watching the probe closely.
“I think HUD has an obligation to look at all platforms to ensure that these algorithms don’t allow, or even accelerate, discrimination,” Sen. Brian Schatz (D-Hawaii) told The Hill. “The ability to target ads is essential for their business models but if they’re targeting ads in such a way that prevents people from getting a public service, that’s not just wrong, it’s against the law.”
The Congressional Black Caucus, which for years urged regulators to look into whether Facebook’s ad-targeting practices were discriminatory, told The Hill that it is “monitoring” HUD’s probes into Twitter and Google.
“We believe all companies should be compliant with any federal law that combats and prevents any form of discrimination,” said Gabrielle Brown, spokeswoman for the caucus, in a statement.
The agency’s moves surprised many Trump administration critics who say officials have not done enough on civil rights enforcement. And for tech companies, it opened a new front for an industry already under immense scrutiny from Washington.
But watchdog groups welcomed the actions and noted that alleged discrimination is not a new problem for the tech industry.
Facebook in recent years has seen a barrage of lawsuits and Equal Employment Opportunity Commission complaints from civil society groups claiming the social media giant allowed advertisers to exclude or target ads based on users’ gender, age, race, national origin, familial status and more.
In response, Facebook earlier this month announced sweeping changes to its targeted advertising practices in housing, employment and credit, three industries protected under federal civil rights laws.
Advertisers dealing with housing, employment and credit will now be held to different standards on Facebook. They will not be allowed to target users based on age, gender or zip code, and they will have far fewer microtargeting options that could be used as proxies to exclude protected classes.
“With Google and Twitter, we definitely would see similar concerns,” Miranda Bogen, a senior policy analyst with digital civil rights group Upturn, told The Hill. “Any ad platform that allows this sort of ad-targeting, or that tries to optimize delivery to the right kind of people, will have to grapple with the same sort of issues that Facebook has been called out for.”
Twitter allows advertisers to target users based on their gender and zip code, and it does not treat housing, employment and credit opportunities differently.
The company prohibits ad targeting based on categories including racial or ethnic origin, negative financial status, health, political affiliation, religious background, sexual orientation, gender identity and more.
But it offers targeting options based on users’ familial status, including whether they are single or have children. The Fair Housing Act prohibits discrimination based on familial status, including against families with children under 18.
“That’s something regulators have been attentive to,” Bogen said. “It’s not appropriate to exclude parents with kids from certain types of housing.”
If users see an advertisement that could violate the law on Twitter, they are encouraged to report it to the company, but civil rights lawyers argue that is not enough.
“There is not a distinction made on those platforms that makes clear that targeting in that manner is unlawful if the advertisements are for opportunities related to employment, housing or credit,” Galen Sherwin, who helped settle the Facebook case with the American Civil Liberties Union, told The Hill. “It’s left to the advertisers to regulate themselves.”
Sherwin and Bogen say there are a number of other, less obvious ways that targeted ads can enable discrimination. For example, they highlighted concerns about using zip codes for targeted ads, which could be a proxy for race due to the country’s legacy of racial segregation. Both Twitter and Google offer advertising based on geographic location.
Google allows advertisers to target users according to their online activity on its platform, particularly their search terms, as well as their location.
“Browsing behaviors, and the similarities between them, could end up creating a demographically homogeneous audience without anyone articulating that that should happen,” Bogen said.
Google and Twitter also offer advertisers a “lookalike” targeting option that targets ads at users that are deemed similar to the advertiser’s intended audience.
HUD is taking issue with Facebook’s own lookalike targeting option, claiming it “uses the protected characteristics of people to determine who will view ads regardless of whether an advertiser wants to reach a broad or narrow audience.”
Asked about the HUD probe, a Google spokesman said the company has policies that “prohibit targeting ads on the basis of sensitive categories like race, ethnicity, religious beliefs, disability status, negative financial standing, etc.”
Google does have specific policies around advertising financial products and services, including barring payday loan advertisements, but it does not provide many specifics around its treatment of employment, housing and credit opportunities.
Google and Facebook together accounted for 58 percent of the $111 billion digital ad market in 2018, with Twitter having a much smaller footprint.
Deirdre Mulligan, a professor at University of California, Berkeley who has researched ad discrimination, told The Hill that HUD is raising questions about whether these companies “have a responsibility to think through the ways in which their algorithms on their own might lead to biases in who receives ads.”
There are also many questions about the ad practices.
Google, Twitter and Facebook have not disclosed how many housing, credit and employment advertisements are on their platforms each year. And none of the companies share detailed information about ad targeting by demographics, posing a significant challenge to researchers and regulators.
Negotiations between HUD and Facebook during the investigation broke down after the agency requested “sensitive user data,” Facebook said in a statement following the charges. Similar issues would likely arise during any HUD probe into Google’s and Twitter’s ad-targeting practices.
Industry watchers say tech companies should be on high alert. And lawmakers are already calling on regulators to expand their investigations.
Sen. Richard Blumenthal (D-Conn.) said “big tech” should not be shown “leniency” when it comes to the landmark Fair Housing Act, which HUD claims Facebook violated.
“For too long, big tech has been given a pass on a lot of legal enforcement that applies to other industries,” Blumenthal told The Hill.
“Maybe HUD should be looking at other companies if there are credible allegations.”
