Facebook has told researchers at New York University to stop using a digital tool that tracks how people are targeted with political ads ahead of the Nov. 3 election.
The demand, sent last week and confirmed by NYU on Friday, centers on the academics’ use of a web browser plug-in that gives Facebook users a way to share specific political ads they are seeing on the site.
Political advertisers primarily target their ads to specific demographic groups, so the NYU tool — which collects roughly 16,000 ads each week — allows researchers to see how campaigns and other groups are crafting messages to voters based on race, age, location or other criteria.
In its notice to NYU, Facebook said that the use of the plug-in broke the company’s terms of service, and ordered the academics to stop using the tool by Nov. 30 or face “additional enforcement action.” NYU said it will not take it down.
“We’re not going to comply with it,” Laura Edelson, an NYU researcher who is part of the project, told POLITICO. “What we are doing is perfectly legal and is in the public interest.”
The standoff comes as social media giants are facing increased scrutiny over how political operatives are using their platforms to reach voters online, often in messages that use people’s personal data to pinpoint them with individualized ads.
The plug-in is part of a broader project from the university aimed at archiving all political ads from official campaigns and partisan groups that have spent hundreds of millions of dollars, collectively, to promote Republican and Democratic candidates in the 2020 election.
In that separate relationship with Facebook, NYU researchers can collect some granular information on political ads. The system went down for several hours on Oct. 22 because of a technical glitch, though Facebook and researchers said the blackout was not related to the academics’ plug-in.
In a statement, Facebook said its executives had told the NYU researchers several months ago that the plug-in would violate the company’s policies, and that its own political ad database provided transparency on how groups were buying ads.
Facebook has invested heavily to provide greater transparency around political ads, and has forced those looking to run political ads in the run-up to the election to jump through more hoops, including providing official IDs, before they can buy these paid-for messages.
“In the U.S., there are a few extra steps that advertisers need to take before being able to run any ad about social issues, elections or politics,” Susan Schiff, a Facebook product manager working on its transparency tools, told reporters last month.
Facebook’s transparency push requires advertisers to disclose their affiliations so voters can see who is funding the political ads they see. It also aims to stop groups and potentially foreign governments from targeting would-be voters with false messaging and other harmful content like incitements to violence around the upcoming election.
NYU’s plug-in has already caught several ads that did not have the appropriate disclaimers that are required under Facebook’s community standards. That includes a Latino news and entertainment site running paid-for messages in favor of Joe Biden, as well as separate ads bought by Chinese state media and The Daily Wire, a conservative website.
In recent weeks, the Trump campaign has peppered female voters in a number of crucial battleground states like Florida and Michigan with paid-for messages associating Biden with violent protests across the country allegedly connected to the so-called antifa movement, according to an analysis of Facebook’s ad transparency data by POLITICO. Men in those same swing states have received far fewer of those messages than women.
POLITICO’s analysis of the ads depends partly on information gathered through NYU’s access to the company’s political ad database, but not its use of the plug-in.
The Nov. 30 deadline means no action would be taken against NYU before Election Day or its immediate aftermath. Facebook is banning all political ads from Election Day until a yet-to-be determined date, so it’s unclear what the impact would be after the deadline passes.
This is not the first time that Facebook has taken action against such browser plug-ins, which permit researchers to collect more detailed data on how people are being targeted than what is currently offered through the company’s own transparency tools.
Last year, for instance, a similar tool created by ProPublica was blocked from accessing Facebook’s systems on the grounds that it broke the company’s terms of service by potentially breaching people’s online privacy rights.
Sam Jeffers, the co-founder of WhoTargetsMe, a separate browser plug-in that also allows people to share how they are being targeted on Facebook, told POLITICO that Facebook had yet to contact him about his plug-in.
Edelson, the NYU researcher, said her tool is particularly crucial to understanding how different political groups are targeting voters in the 2020 election, which is being fought online more than previous presidential elections because of the Covid-19 pandemic.
“When things get through the cracks and vulnerable communities are targeted with harmful content, they have a right to know,” she said. “People are concerned about how they are being targeted, and Facebook has not done a good job in providing that level of transparency.”