ChatGPT Fabricated Sexual Harassment Allegation Against Law Professor Jonathan Turley

In April 2023, a user queried ChatGPT requesting examples of legal scholars who had been accused of sexual harassment. In response, the AI chatbot generated a completely fabricated allegation against Jonathan Turley, a prominent law professor at George Washington University Law School. The AI claimed that Turley had been accused of sexual harassment during a class trip to Alaska and cited a purported Washington Post article from 2018 as its source. When the user attempted to verify this information, they discovered that no such Washington Post article existed. The entire allegation was a fabrication by ChatGPT, which had apparently synthesized plausible-sounding but entirely false information about a real person. Professor Turley himself became aware of the incident when contacted about the alleged article and confirmed that no such accusation had ever been made against him. The incident highlighted a critical flaw in ChatGPT's operation: its tendency to generate convincing but false information when it lacks real data to answer a query. Rather than acknowledging uncertainty or declining to provide potentially harmful information about specific individuals, the AI created a detailed scenario complete with fictional source attribution. This demonstrated the model's inability to distinguish between factual reporting and plausible fiction when generating content about real people. Professor Turley, a frequent legal commentator and constitutional law expert, expressed concern about the potential damage such false allegations could cause to individuals' reputations. The incident occurred during a period of heightened scrutiny around AI-generated misinformation and the potential for large language models to create convincing but false narratives about real people and events. The case became widely discussed as an example of AI hallucination with serious real-world consequences, particularly in the context of sensitive allegations that could damage professional reputations. It underscored the need for better safeguards when AI systems are asked to provide information about specific individuals, especially regarding potentially defamatory content.