GPT-5 has received a rather mixed reception so far. For some, it’s been a significant upgrade across the board. For others, it’s been a let-down, and if anything, it made the tool worse. OpenAI is now trying to fix that with a variety of small changes and upgrades. The latest of which is to make the model warmer and friendlier.
In a post on X, OpenAI stated: “We’re making GPT-5 warmer and friendlier based on feedback that it felt too formal before. Changes are subtle, but ChatGPT should feel more approachable now.”
We’re making GPT-5 warmer and friendlier based on feedback that it felt too formal before. Changes are subtle, but ChatGPT should feel more approachable now.You’ll notice small, genuine touches like “Good question” or “Great start,” not flattery. Internal tests show no rise in…August 15, 2025
OpenAI has often struggled to get the tone of its model correct. In the past, it has been made far too sycophantic, overly praising users and sometimes telling them they were correct, even if they were wrong.
The post goes on to say: “You’ll notice small, genuine touches like “Good question” or “Great start”, not flattery. Internal tests show no rise in sycophancy compared to the previous GPT-5 personality. Changes may take up to a day to roll out.”
It’s not immediately clear exactly how much has changed here, but OpenAI is clearly struggling to get the chatbot’s tone nailed down.
GPT-5 offered a variety of customisation options for the model’s personality. You can make it more chatty, blunt, witty or reply with a certain style to its writing. But OpenAI is also trying to make the default voice more likeable too.
The success of this is hard to measure, but you can see how each chatbot has it’s own style Anthropic’s Claude, for example, is more straightforward and with fewer ‘personality’ traits. But OpenAI prefers ChatGPT to be ‘friendly,’ which comes with its own challenges.
The internet’s response
Where OpenAI once enjoyed seemingly endless support on its decisions, that seems to be changing. The replies to the company’s post on X involved a lot of outcry and one common opinion; that GPT-5 isn’t what they were promised.
“This is not at all what people want. I do not want it to pretend to be human, I want it to dispense information parsimoniously [ed’s note: efficiently]. Please do not force the silent majority to suffer due to the predictions of the vocal minority,” said X user Mark Valorian.
Another user, Gosucoder, followed up with “No, please don’t do this, what this change does is just reinforce that GPT-5 is more for people that build attachment to AI instead of using it as a tool.”
The launch of GPT-5 certainly hasn’t gone as smoothly as OpenAI would have liked. This is just one of many concerns users have raised about the tool. But maybe, this latest update will do enough to ease those worries people have about running an over-friendly ChatGPT.
More from Tom’s Guide
Back to Gaming Laptops
