OpenAI’s ChatGPT is in the spotlight after it reportedly made false sexual harassment allegations against a prominent lawyer. Jonathan Turley, Professor of Law at George Washington University, was sent an email by another lawyer stating that his name had been included on a list of “legal scholars who have sexually harassed someone” supposedly generated by the AI chatbot. Although the email said the source was a Washington Post article, the Post reported that the article did not exist.
When contacted by Insider, Turley called the claims “chilling” and noted that he had never been accused of sexual harassment or assault. OpenAI responded that when people sign up for ChatGPT, the company does its best to make it clear the chatbot may not always generate accurate answers. It then added that improving factual accuracy is a major focus for the company.
Incidents like this demonstrate the potential for language models, such as OpenAI’s GPT-4 and Microsoft’s Bing chatbot, to generate falsehoods and nonsensical speech. Such models are typically based on data from Reddit and Wikipedia, which are not usually fact-checked.
Researchers of AI and experts in the field have stressed the need for major progress in the accuracy and accountability of these models. Kate Crawford, Professor at the University of Southern California at Annenberg and research at Microsoft, spoke of the incident involving Turley and asserted that such accusations based on AI chatbot responses are concerning.
OpenAI is an artificial intelligence research laboratory co-founded by entrepreneur Elon Musk. The company is best known for its Machine Learning models, which are capable of generating human-like natural language processing capabilities. OpenAI was founded in the belief that artificial general intelligence is the most important and pressing challenge of the 21st century. Through its research, OpenAI’s mission is to ensure that artificial general intelligence is developed safely and responsibly.
