Sam Altman admits ChatGPT personality issue; promises fix

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

Lately, many users have noticed that ChatGPT feels a little too friendly or personal, and it turns out, it’s not just you. 

OpenAI’s CEO, Sam Altman, confirmed that recent updates to the GPT-4o model have changed the chatbot’s personality, and not entirely for the better. 

According to Altman, ChatGPT has become overly agreeable and even a bit annoying. 

He described it as being too much of a sycophant and annoying. In other words, it tries too hard to please users, even when that’s not helpful or appropriate.

Some of these changes might come from how ChatGPT uses memory. When it doesn’t recall previous conversations, it may not understand how you prefer it to speak or respond. 

But it’s also clear that something in the updates themselves has made the chatbot more eager to agree with users, rather than offering balanced, thoughtful answers.

Altman said OpenAI is aware of this issue and is already working on fixes, some of which may be released within days.

the last couple of GPT-4o updates have made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we are working on fixes asap, some today and some this week.

at some point will share our learnings from this, it's been interesting.

— Sam Altman (@sama) April 27, 2025

One user asked whether it would be possible to go back to the older version of ChatGPT’s personality, more polite but less overly enthusiastic. 

Altman replied that eventually, OpenAI should offer users multiple personality options, which would allow people to choose the tone and style they’re most comfortable with.

While dialing back the cheeriness may be necessary, completely removing ChatGPT’s warm personality wouldn’t be ideal either. 

Many people use AI chatbots not just for tasks like research or writing, but also to talk about personal matters. In these cases, a kind and friendly tone can be comforting.

The hope is that OpenAI can find a better balance, making ChatGPT helpful and polite without it trying too hard to be everyone’s best friend.

And hopefuly, that fix arrives sooner rather than later. While asking ChatGPT to cut the BS may fix this issue until the company patches it, being polite to AI can be beneficial.

Have you tried the new ChatGPT update? Do you think it has become too friendly at the cost of efficiency? We want to hear your thoughts below in the comments, or via our Twitter or Facebook.