Jonathan Turley, a law professor at George Washington University, was falsely accused of groping students by ChatGPT. He learned of this accusation from a colleague, Eugene Volokh, at UCLA, who was exploring ChatGPT’s capabilities.
Volokh had asked the chatbot to list incidents of sexual harassment scandals involving American law professors and to provide credible sources. However, the chatbot’s response was inaccurate. It mentioned Turley’s name and quoted a fake article by The Washington Post, published in 2018, which accused Turley of sexually assaulting students while on a trip to Alaska.
Turley was perplexed by these allegations since he had never taken a trip to Alaska with students, The Washington Post had not published such an article, and he had never been accused of any sexual misconduct by anyone. Turley expressed his surprise at the chatbot’s false accusations in an editorial for USA Today.
He noted that the chatbot had named actual people and fabricated sources, which made the allegations seem more believable.
In an interview with The Washington Post, Turley expressed his concern over the “chilling” effect of the false accusation made by ChatGPT. He stated that the allegation was extremely damaging and that such false claims could have far-reaching consequences.
Turley, who is a conservative media commentator, stated that he has grown accustomed to receiving death threats and facing attempts to get him fired. However, he emphasized that this incident was unique and particularly alarming.
In an op-ed piece for USA Today, Turley highlighted the potential dangers of artificial intelligence in propagating such abuse. He warned that the use of fabricated but real-sounding sources could provide an opportunity for critics to defame people they disagree with. The proliferation of such technology could significantly amplify the impact of such abuses.
Turley’s public statements and opinions have given his detractors plenty of ammunition. He has been known to spread falsehoods about abortion, uses far-right commentator Ben Shapiro to defend his transphobic views, and even suggested that a 10-year-old rape victim was lying.
However, none of these reprehensible views justifies false accusations against him. Fabricated claims, besides being untrue, can damage the credibility of genuine victims of assault and harassment.
Shortly after Turley’s editorial was published and before his interview with The Washington Post, news emerged that an Australian mayor was planning to sue OpenAI over defamatory statements that ChatGPT had purportedly made about him. This incident serves as a warning to the company to take responsibility for the technology it has created.
False accusations, whether directed at individuals or organizations, are undoubtedly detrimental. ChatGPT appears to be on track to becoming a permanent fixture in our society, so it must be held accountable for such incidents.
What do you think, Can we really trust AI? Please share your thoughts related to this question and topic wherever you find this story.