OpenAI has recognised the problems with GPT-5 and made it more humane
OpenAI has decided to slightly "clean up the character" of its new GPT-5 model. On Friday, the company announced that it was updating the artificial intelligence, making it "warmer and friendlier".
The GPT-5 release didn't turn out to be as smooth as planned. The company's CEO Sam Altman admitted that the launch went "a little more nervously than we expected." Many users directly said that they liked the previous GPT-4o better - say, the new interlocutor sounds too dry and detached.
Here's What We Know
Now OpenAI has added a small "human" touch to GPT-5: phrases like "Good question" or "Great start". According to the developers, these are not flattering compliments, but an attempt to make the tone of communication more natural. The company claims that internal tests have not shown an increase in "obsequiousness" in the model compared to the last version.
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.
- OpenAI (@OpenAI) August 15, 2025
You'll notice small, genuine touches like "Good question" or "Great start," not flattery. Internal tests show no rise in...
At a dinner with reporters, OpenAI executives tried to shift the focus to post-GPT-5 plans, but the "elephant in the room" - the botched launch - was still the main discussion. Vice President Nick Turley explained that the model was "too straightforward" before the update, but now it will become warmer in communication.
In other words, the GPT-5 has officially taken a "courtesy course." But whether users will like the new version is still an open question.
Source: TechCrunch
