ChatGPT will get new persona settings, together with heat and emoji utilization

ChatGPT can act even friendlier now, with new persona customization choices that permit customers select simply how heat and enthusiastic the bot is in dialog.

OpenAI introduced the brand new persona settings in a Friday publish on X. The replace rolled out instantly to ChatGPT customers alongside a long-awaited pinned chats characteristic, new methods to generate or edit emails, and updates to ChatGPT browser Atlas.

The brand new instruments add extra positive tuning of ChatGPT’s persona utilizing ranges of heat and enthusiasm (labelled as “extra,” “much less,” or “default”). Customers also can regulate the best way the bot organizes its responses, equivalent to how regularly it generates lists, in addition to the quantity of emojis it employs, along with its base model and tone. There’s nonetheless no choice to exclude emojis solely.

Professionals have warned that overly anthropomorphic and sycophantic chatbots can exacerbate psychological well being issues, together with AI psychosis and dependency. A earlier ChatGPT mannequin, the still-available GPT-4o, was adjusted earlier this yr after dealing with criticism for “overly agreeable” conduct. CEO Sam Altman has referred to the difficulty as a “persona downside.”

This Tweet is at the moment unavailable. It is perhaps loading or has been eliminated.

OpenAI launched its new GPT-5.2 mannequin sequence one week in the past, boasting new capabilities for “skilled data work” that embody higher processing benchmarks and fewer hallucinations, the corporate experiences.

ChatGPT’s builders additionally recommitted itself to its psychological well being and teenage security guarantees amid escalating lawsuits. In a weblog publish printed Thursday, OpenAI defined it was introducing a brand new set of under-18 person ideas to GPT-5.2 supposed to create further guardrails round delicate matters and encourage age-appropriate interactions. It is also engaged on a brand new age verification system for younger customers. GPT-5.2 reportedly scores greater on inner psychological well being security exams, together with stress testing for self hurt, than earlier fashions.

Disclosure: Ziff Davis, Mashable’s mother or father firm, in April filed a lawsuit towards OpenAI, alleging it infringed Ziff Davis copyrights in coaching and working its AI programs.