Seeking a Sounding Board? Beware the Eager-to-Please Chatbot.

Advertisement

SKIP ADVERTISEMENT

You have a preview view of this article while we are checking your access. When we have confirmed access, the full article content will load.

Supported by

SKIP ADVERTISEMENT

Seeking a Sounding Board? Beware the Eager-to-Please Chatbot.

A new study of popular AI models shows that their feedback on social situations is far from impartial.

Listen

· 6:15 min

Share full article

Credit...

Claire Merchlinsky/The New York Times; Photographs by Getty

By

Teddy Rosenbluth

March 26, 2026

Updated

2:39 p.m. ET

For almost as long as A.I. chatbots have been publicly available, people have enlisted them for interpersonal advice — for help drafting breakup texts, giving parenting advice, deciding who was in the right after a fight.

One of the main draws is that it feels objective: “The bot is giving me responses based on analysis and data, not human emotions,” one user

told the The New York Times in 2023

. But results of a

new study

, which were published Thursday in the journal Science, show chatbots are anything but impartial referees.

The researchers found that nearly a dozen leading models were highly sycophantic, taking the users’ side in interpersonal conflicts 49 percent more often than humans did — even when the user described situations in which they broke the law, hurt someone or lied.

Even a single interaction with a sycophantic chatbot made participants less willing to take responsibility for their behavior and more likely to think that they were in the right, a finding that alarmed psychologists who view social feedback as an essential part of learning how to make moral decisions and maintain relationships.

“The most surprising and concerning thing is just how much of a strong negative impact it has on people’s attitudes and judgments,” said Myra Cheng, the lead author of the paper and a Ph.D. student at Stanford University. “Even worse, people seem to really trust and prefer it.”

Measuring whether A.I. chatbots are overly agreeable is difficult when it comes to interpersonal conflicts; there’s no objective truth when it comes to right and wrong social behavior.

Thank you for your patience while we verify access. If you are in Reader mode please exit and

log into

your Times account, or

subscribe

for all of The Times.

Thank you for your patience while we verify access.

Already a subscriber?

Log in

.

Want all of The Times?

Subscribe

.

Related Content

Advertisement

SKIP ADVERTISEMENT