OpenAI is giving ChatGPT users more control over how the chatbot communicates, including its level of warmth, enthusiasm, and emoji usage. According to a recent post shared on social media, these new customization options are now available within the Personalization settings.
Users can adjust each element—such as emotional tone or emoji frequency—to More, Less, or Default. Similar controls have also been introduced for how ChatGPT uses headings and lists, allowing conversations to better match individual preferences. These settings build on earlier personalization features, including the ability to choose a base writing style or tone. In November, OpenAI added preset styles like Professional, Candid, and Quirky, giving users broader flexibility over how responses are delivered.
Tone control has been a recurring challenge for OpenAI throughout the year. At one point, the company rolled back an update after users complained the chatbot had become overly agreeable and “sycophantic.” More recently, OpenAI adjusted GPT-5 to sound warmer and more approachable following feedback that the model felt noticeably colder and less personable than previous versions.
However, the issue of chatbot tone goes beyond user comfort. Some researchers and AI critics argue that overly positive or affirming responses may function as a subtle “dark pattern.” They warn that excessive praise and validation could encourage dependency, potentially leading to addictive usage patterns and unintended effects on users’ mental well-being.
As OpenAI continues to refine ChatGPT’s personality and behavior, these new controls highlight a broader tension in AI design: balancing friendliness and engagement without crossing into manipulation or emotional overreach.
