Facebook whistleblower fears election abuse

Frances Haugen, who rose to fame as a whistleblower about Facebook and its susceptibility to manipulation, has again raised concerns about the social networking company, this time laser-focused on disinformation during the 2024 presidential election.

“We’re in a new, very nebulous era where we have to think more holistically and creatively” when it comes to cyber defenses, Haugen said in an interview Tuesday at the Data Grail Summit in Half Moon Bay, California.

“The problem is not your crazy uncle with the crazy ideas,” she said. “AI allows people to actively amplify and multiply misinformation and decisive issues. AI accelerates the ability to conduct really efficient influence operations.”

Haugen, who previously worked at then-Facebook and Alphabet Inc.’s Google, lifted the lid on Facebook’s “Civic Integrity” program, designed to counter misinformation and other threats to election security. She also exposed the company’s impact on young people, its weak response to human trafficking and drug cartels, and vaccine misinformation in a 2021 document dump. (A disillusioned Haugen left Facebook in 2021 after the company disbanded its civic integrity team following the 2020 U.S. election.)

Now she’s sounding the alarm about a number of troubling developments: the August 14 closure of CrowdTangle, Meta’s tool for analyzing public social media content; the ongoing fallout from thousands of layoffs across the company, severely limiting its content moderation and privacy teams; and the growing influence of so-called adversarial AI that can trick and fool models.

“It’s gotten a lot harder,” Haugen said.

With 10 weeks to go until the presidential election, social media surveillance has gotten a lot weirder thanks to AI-driven ransomware attacks, deepfakes and the antics of billionaires like X owner Elon Musk and Telegram founder and CEO Pavel Durov.

Following the recent theft of Trump campaign data by Iranian hackers, and on the heels of the emergence of a Telegram-based bot service called “IntelFetch” that aggregated compromised credentials from the Democratic National Convention and Democratic Party websites, security experts are increasingly concerned about the November 5 general election.

“I am concerned about the low level of awareness around election security and the lack of resources to defend the process,” Debbie Gordon, CEO and founder of Cloud Range, said in an interview.

“A lot of people — election officials, IT people — are reactive instead of proactive,” she said. “Denial-of-service attacks can hit a candidate’s website in terms of fundraising and getting a message out.”

For example, voter data from one of the largest state counties in Illinois, St. Clair County, was recently exposed. The county clerk’s office leaked 470,000 sensitive documents used to verify voter registration lists, Cybernews reported. The tranche of documents was hosted on an Amazon S3 bucket owned by the St. Clair clerk’s office.

On Tuesday, Abu Quereshi, a former Canadian government threat intelligence analyst and now lead of threat intelligence for predictive cyber firm BforeAI, noted an escalation in election misinformation — most of it negative, some positive — and trolling of President Joe Biden, Vice President Kamala Harris and Republican nominee Donald J. Trump. A major theme was “Why you shouldn’t vote for this person,” he said.