According to OpenAI’s Chief Technology Officer Mira Murati in an interview with Time magazine, the bot “may make up facts” as it composes statements.
According to a report in the Business Insider, Mira called this a “central problem.” According to her, ChatGPT creates its responses by anticipating the logical word that should come next in a sentence, although what the bot deems logical may not necessarily be correct.
Users may take care of that by continuing to interact with the bot & challenging answers they don’t think are accurate, she said.
In the same interview the CTO stated that artificial intelligence tools require regulatory oversight.