ChatGPT now provides options for personality customization, enabling users to modify the bot’s warmth and enthusiasm during interactions.

OpenAI revealed these new features in a post on X last Friday. The update was rolled out immediately to ChatGPT users, along with a pinned chats feature, innovative methods for composing or altering emails, and enhancements to the ChatGPT browser Atlas.

These tools permit users to fine-tune ChatGPT’s personality by changing the warmth and enthusiasm levels, designated as “more,” “less,” or “default.” Users can also alter how the bot structures its responses, including the frequency of lists and the integration of emojis, in addition to its fundamental style and tone. However, users cannot fully eliminate emojis.

Experts have warned that highly anthropomorphized and sycophantic chatbots may exacerbate mental health problems, such as AI-related psychosis and dependency. A prior version of ChatGPT, GPT-4o, underwent modifications earlier this year following criticism for being “excessively agreeable.” CEO Sam Altman has characterized this as a “personality issue.”

OpenAI debuted its new GPT-5.2 model series a week ago, boasting enhanced capabilities for “professional knowledge work,” including improved processing metrics and reduced inaccuracies, as reported by the company.

The developers of ChatGPT have also reiterated their dedication to mental health and the safety of teenagers amid increasing lawsuits. In a blog post shared on Thursday, OpenAI introduced a new set of principles for users under 18 for GPT-5.2, designed to establish additional safeguards around sensitive subjects and encourage age-appropriate interactions. The company is also working on a new age verification system for younger users. GPT-5.2 is said to perform better on internal mental health safety evaluations, including stress tests for self-harm, compared to earlier models.

Disclosure: Ziff Davis, the parent organization of Mashable, initiated a lawsuit against OpenAI in April, claiming it violated Ziff Davis copyrights while training and operating its AI systems.