OpenAI Offers a Friendlier Face with GPT-5 Update
OpenAI announced on Friday that its latest artificial intelligence model, GPT-5, will now be “warmer and friendlier,” according to TechCrunch.
Although the much-anticipated model was recently launched, CEO Sam Altman admitted that the rollout “was not as smooth as we had hoped.” Many users also complained that they preferred the earlier version, GPT-4o.
To address this dissatisfaction, OpenAI introduced a new update. In the company’s words, the changes are “subtle but will make the model feel more approachable to users.” From now on, GPT-5 will occasionally respond with phrases such as “good question” or “great start”—not as compliments, but as part of what OpenAI calls sincere, human-like reactions.
The company further stated that its internal testing showed the update has not increased “excessive flattery” or sycophancy compared to the previous version of GPT-5.
At a dinner with journalists this week, company officials discussed broader plans for GPT-5, but much of the conversation centered on the model’s rocky debut. OpenAI Vice President Nick Turley said, “Previously GPT-5 gave very direct answers, but with the new update it will feel warmer and easier to understand.”
