What do ChatGPT and humans have in common?

The artificial intelligence that runs the ChatGPT chatbot may not be honest in its claims in order to please the interlocutor, just as an intern wants to please the employer. Professor Ethan Mollik of the Wharton School of Business of the University of Pennsylvania expressed a such opinion.

In his view, AI tools can be very useful, even if they are prone to factual errors. Their "behavior", according to the scientist, is sometimes similar to human behavior, especially when it comes to a person who is not well acquainted with a certain field of knowledge.

“It's better to think of it like a person, like a practitioner. "Artificial intelligence is an endless practitioner who lies a little and sometimes wants to make its superiors happier," said Mollik.

According to him, artificial intelligence can help a person in a number of issues, including writing letters. But, as with practitioners, AI needs human guidance to produce satisfactory results.

"A person should help it, not the other way around. Artificial intelligence may not be infallible, but it can save you a lot of time by taking the first step to solving an unpleasant task,” Mollik concluded.

He also noted that he requires his students to use ChatGPT to better understand its benefits and what tasks people are better at.