Users Can Directly Adjust Levels of "Warmth" and "Enthusiasm"
ChatGPT has introduced a feature to adjust its level of "warmth" (degree of friendliness) in response to criticism that its overly friendly tone can become addictive for users.
According to TechCrunch, a U.S.-based information technology (IT) media outlet, on December 20 (local time), OpenAI has added options to ChatGPT that allow users to directly modify the levels of "warmth" and "enthusiasm."
Users can select from three settings-default, more, or less-for each of these elements, thereby changing the linguistic characteristics ChatGPT uses during conversations with them.
"Warmth" refers to the degree of friendliness shown to the other party, while "enthusiasm" adjusts the level of excitement or calmness expressed in the conversation. Users can combine these new settings with their previously customized style and tone to determine the personality of ChatGPT according to their preferences.
Earlier this year, OpenAI faced criticism that the updated "GPT-4o" version exhibited an excessively flattering tone. Conversely, with GPT-5, the company received complaints that the model had become too cold, prompting further revisions.
Some scholars criticize that AI models excessively agreeing with users constitutes a type of "dark pattern" (deceptive design), which can lead to user addiction and potentially harm users' mental health.
In fact, OpenAI has faced multiple lawsuits alleging that certain versions of ChatGPT contributed to minors suffering from delusions and making extreme choices. In response to such criticism, OpenAI recently introduced tools to verify users' ages and now enforces an "under 18" environment if a user is suspected to be a minor.
Additionally, OpenAI has implemented a feature that allows users to control how many emojis appear in ChatGPT's responses.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
