New AI tools can dig up skeletons in politicians’ social media histories—and yours, too

Character, as they say, is what you do when no one’s watching. What’s been hard to ignore during the current election, however, is that character also seems to be what you do when only certain people are watching. Or when you think only certain people are watching, anyway.

Ever since JD Vance became Donald Trump’s running mate back in July, the Ohio Senator has been haunted by previous online activity intended for select audiences. Between unflattering clips from old podcasts and some unsavory connections within the financial app Venmo, the first millennial to land on a presidential ticket is also the first VP pick whose web presence has come back to bite them like an edgy comedian hired on Saturday Night Live.

JD Vance’s rocky candidacy is proof that, no matter your political ideology, being terminally online has become risky for anyone entering the public eye in a high-profile gig. Not only has it gotten harder to hide the digital skeletons lurking in one’s closet, but the professional skeleton-finders are doing more thorough closet-inspections than ever—especially in the realm of politics.

Sifting through your internet history

The state of being terminally online involves logging on for so long within one’s preferred echo chamber that it informs practically all your opinions and makes you comfortable sharing them widely and confidently. In a political environment where every awkward conversation has the potential to go viral, this can be a problem. It’s a fairly recent problem.

JD Vance and his fellow millennials were only in their early-20s when Facebook launched in 2004. Since then, they’ve accrued 20 years worth of FB posts, tweets, Livejournal entries, and YouTube vlogs; all while a culture of “this you?-style” receipts emerged, teaching onlookers how to take a magnifying glass to digital footprints.

The problem isn’t just limited to the folks whose names are actually on the ballots either. For instance, AI-powered screening company Ferretly has just launched a social media vetting platform specifically for political staffers.

“I don’t believe in censorship, but I think what someone says publicly can have consequences,” says Ferretly founder Darrin Lipscomb. “And I think if they want to join a company and be part of the culture of that company, that company has every right to assess those public statements.”

Ferretly has been letting companies know what would-be employees have said online in public spaces since Lipscomb founded it in 2019. The service uses AI to scan social media and other publicly available data, going back 10 to 15 years, to root out potential risks and behaviors beyond the scope of traditional background checks—including inflammatory rhetoric, hate speech, bullying, drug use, and nudity.

With customers in 32 countries, the company’s eclectic client base includes both collegiate and professional sports teams, various police departments, and reality TV stars and influencers. One need not look very far to come across problematic hires in these fields, which perhaps could have been flagged by a company like Ferretly.

The company didn’t set out to make a platform expressly for election workforce screening, though.

“We weren’t even thinking about this market or vetting political appointees,” Lipscomb says.

“It started with some specific behaviors we were surfacing from social media around politics, government speech, and extremism, and we went from there.”

After independently noticing a rise in these behaviors among those screened in other fields, Ferretly says it was hired to vet some Member of Parliament candidates in the U.K., on both sides of the aisle. The company eventually launched the service as an Election Workforce platform and is now working with some PACs in the U.S..

The platform, which can be tailored to a client’s specific needs, goes beyond standard screening, with an emphasis on extreme rhetoric and visuals. Its image-based machine-learning classifiers can sift through thousands of social media posts and memes to pick up on offensive gestures like the middle finger, extremist symbols like Nazi insignia and terrorist flags, along with images of guns, sharp objects, and explosives. If a potential campaign manager has posed jauntily next to The Communist Manifesto in a publicly posted photo, or while wearing a QAnon T-shirt, the platform can likely find it, Lipscomb says.

Digging up old tweets

The fact that Ferretly’s screening platform is in demand is only the latest sign of how scrutinized social media has become for anyone seeking political jobs. Beyond social media, though, the platform can also, if a campaign calls for it, search through transcripts of publicly available podcasts in search of various keywords and phrases. (It can’t comb through video yet, but that’s a challenge Lipscomb hopes to tackle soon.) Just imagine what it might turn up for a terminally online political job candidate with fiery opinions about childless cat ladies, were that something a campaign wished to avoid having to address.

As social media vetting becomes more commonplace for high-profile job candidates, people aren’t likely to stop engaging in the kind of behavior the vetting is meant to capture as much as they might learn how to hide it better. Fledgling comedians haven’t stopped creating offensive material, after all; they’ve just learned to nuke as much of their social media presence as possible when they’re in the mix for getting cast on SNL. Some people are even hiring Ferretly to do screenings on themselves, says Lipscomb, in case any blemishes turn up that they might have forgotten about.

Perhaps the key to being terminally online is just knowing when and how to cover your tracks.  

“People can say whatever they want but they should also understand that there’s tools out there, including ours, that companies can use to assess what those statements are,” Lipscomb says. “If you’re not smart enough to realize that, then that’s probably something against you.”