Getty / Futurism
Time's Up

ChatGPT Falsely Accuses Law Professor of Sexual Assault

byNoor Al-Sibai
4.6.23, 12:23 PM EDT
Getty / Futurism

Nobody deserves this.

Phony Bologna

ChatGPT is throwing around some serious allegations.

In an editorial for USA Today, George Washington University law professor Jonathan Turley wrote that he was surprised to learn from a colleague who'd been researching ChatGPT that it had falsely accused him of groping students.

As Turley wrote, that colleague, UCLA's Eugene Volokh, asked the chatbot to describe scandals involving American law professors being accused of sexual harassment and to cite media sources. It obliged, sort of: it named names and made up sources, with one of those "sources" being a phony 2018 Washington Post article that it falsely claimed said Turley had sexually assaulted students during a trip to Alaska.

"It was a surprise to me since I have never gone to Alaska with students, The Post never published such an article, and I have never been accused of sexual harassment or assault by anyone," the GW law professor wrote.

In an interview for an actual WaPo article, Turley said the whole debacle was "quite chilling."

"An allegation of this kind is incredibly harmful," he told the newspaper.

Just Criticism

A conservative media commentator, Turley said that although he's grown used to death threats and people trying to get him fired, this situation was different.

"AI promises to expand such abuses exponentially," he wrote, noting that critics who will often run with anything presented to them could easily use these sorts of real-sounding citations to disparage people they disagree with.

Cards on the table: Turley's very-public opinions do, indeed, give critics a lot to work with. He's lied about abortion, cited far-right commentator Ben Shapiro in his defenses of transphobia, and suggested that a 10-year-old rape victim was lying.

None of that, however, makes him worthy of false accusations, which along with being untrue are also harmful to the credibility of the many, many people who have suffered from assault and harassment.

Just after Turley's editorial came out (and just before his IRL interview with the WaPo dropped), news broke than an Australian mayor was readying the world's first-ever lawsuit against OpenAI for defamatory things that he says ChatGPT had spewed about him — a signal to the company that, if nothing else, it needs to get a hold on its tech.

False accusations, regardless of who they're lobbied against, are unequivocally bad — and if ChatGPT is going to, as all indicators suggest, become a permanent mainstay in our society, it needs to be held accountable on that front, too.

More on AI: Former Google CEO Warns That Humans Will Fall in Love With AIs


Share This Article