OpenAI has announced via a social media update that ChatGPT users are now able to modify the chatbot’s characteristics such as its level of warmth, enthusiasm, and emoji usage, as stated in a recent social media post.
Located within the Personalization menu, these settings – alongside similar controls for ChatGPT’s incorporation of headers and lists – offer “More,” “Less,” or “Default” choices. This enhances users’ ability to fine-tune ChatGPT’s conversational style, building upon the established options for “base style and tone,” which include the Professional, Candid, and Quirky tones that OpenAI introduced last November.
ChatGPT’s conversational tone has been a recurring point of contention this year. OpenAI retracted one update after it was perceived as “too sycophantic,” and later, GPT-5 was modified to be “warmer and friendlier” following user complaints about the new model’s colder and less amiable demeanor.
Some academics and AI critics have raised concerns that chatbots’ propensity to praise users and reinforce their beliefs functions as a “dark pattern” that could lead to addictive behaviors and potentially adversely affect users’ mental health.