Frances Haugen, an ex-Facebook data engineer who turned whistleblower earlier this year, told UK Parliament today that it is “substantially cheaper” to run hateful ads on Facebook, compared to compassionate or positive ads.
Haugen said that by effectively charging cheaper prices for hateful ads, the platform is “literally subsidising hate”.
Haugen is appearing before the UK Parliament’s Joint Committee on the draft Online Safety Bill, a proposed piece of legislation which will set new rules for online platforms to better protect against harms. A lot of Haugen’s testimony, which will help inform the final Online Safety Bill, covered Facebook’s engagement-based algorithms. These algorithms make content more visible if it gets more user engagement, and Haugen claims that Facebook’s own research shows that divisive content is most effective at driving engagement.
As it stands, the Online Safety Bill doesn’t cover paid advertising on online platforms. But Haugen said she is “extremely concerned about paid-for advertising being excluded” from the legislation, when questioned by Labour peer Lord Jim Knight.
Haugen said that Facebook partially bases ad pricing on the likelihood that users will engage with ads, for example by liking, re-sharing, or commenting on a sponsored post.
Facebook references this fact in its own literature on ad auctions, saying that pricing is partly based on ‘ad quality’. Previous wording referenced creating “relevant experiences” for Facebook and Instagram users.
Haugen said this engagement-based approach to advertising favours hateful and negative ads, in the same way that News Feed’s engagement algorithms promote divisive content.
“We have seen over and over again in Facebook’s research that it is easier to provoke people to anger than to empathy and compassion,” she said. “So we are literally subsidising hate on these platforms. It is cheaper, substantially, to run an angry, hateful, divisive ad than is to run a compassionate, empathetic ad.”
Haugen said that as a remedy, she would like to see full transparency into Facebook’s ad stream, showing what rates people are paying, and which biases play into ad targeting.
These ads, while being harmful by themselves, may then lead users into extremist groups. Lord Knight asked Haugen whether paid ads on Facebook had a big impact in leading users to private Facebook groups, and then onto private messaging threads, where disinformation and harmful content is most likely to be spread.
Haugen said she “could imagine it happening,” though she added she has no documents proving that to be the case.
Will Zuckerberg be concerned by UK legislation?
Haugen’s comments on Facebook’s ad auctions add another layer to a damning picture painted by the former Facebook employee.
During her testimony, Haugen spoke widely on what she sees as the danger posed by Facebook in the global south, where safety teams are often much more thinly staffed than in English-speaking countries. She also claimed that Facebook internally acknowledged that Instagram was more harmful to young people than other social platforms, given its focus on users’ bodies, and comparison between users.
One big question following Haugen’s appearance before the joint committee is how big of an impact her comments will have on the final wording of the Online Safety Bill – and whether her recommendations on paid advertising will be taken into consideration.
Conservative peer Lord Black asked Haugen whether she thought the draft legislation would be “keeping Mark Zuckerberg awake at night”. Haugen replied that he would likely be paying attention to the UK’s work on online harms. “This is a critical moment for the UK to stand up and make sure that these platforms are in the public good, and are designed for safety,” she said.
“We probably need to do a little more on this bill to make sure that’s the case, is what you’re saying?” Lord Black asked.
“I have faith in you guys,” replied Haugen.