ChatGPT learns some manners (you can turn them off)

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.

OpenAI has given ChatGPT a fresh suit with its latest GPT-5.2 update, meant to make it more polished and capable, but not everyone is thrilled with how it sounds while doing it.

Some users say it feels curt. Others say it’s weirdly sassy. And a few insist it’s just judging them.

Luckily, OpenAI has decided the problem isn’t the AI, it’s vibes. And vibes, apparently, are customizable now.

In a post on X, OpenAI announced a new set of personality sliders tucked into ChatGPT’s Personalization settings. 

Users can now tweak how warm, enthusiastic, emoji-happy, or list-obsessed their chatbot is. 

You can now adjust specific characteristics in ChatGPT, like warmth, enthusiasm, and emoji use.

Now available in your "Personalization" settings. pic.twitter.com/7WSkOQVTKU

— OpenAI (@OpenAI) December 19, 2025

Each option comes with a simple choice: more, less, or default. In other words, you can turn ChatGPT into a cheerful helper, a neutral assistant, or something closer to a very competent instruction manual.

These new knobs join the “Base style and tone” controls introduced last month with GPT-5.1, which let users pick between Professional, Candid, or Quirky. 

Together, they form what is essentially a personality mixing board for ChatGPT, because nothing says “cutting-edge AI” like arguing over how many exclamation points it should use.

The move is also OpenAI doing a bit of damage control. Earlier this year, when GPT-5 rolled out to replace GPT-4o, users quickly noticed something was off. 

The new model was smarter, sure, but also colder, less conversational, and more “corporate email reply” than friendly assistant. 

Complaints rolled in, memes followed, and OpenAI responded by letting users switch models and promising to make GPT-5 feel warmer over time.

GPT-5.2 is the latest step in that promise. Instead of guessing how friendly users want their AI, OpenAI is letting them decide. 

Want cozy and enthusiastic? Slide it up. 

Want minimal emojis and straight answers? Dial it back.

With AI getting more capable, people don’t just want correct answers. They want answers that don’t feel like they came from a passive-aggressive coworker.