OpenAI Just Made GPT-5 'Warmer and Friendlier' with a Big Personality Update — and There's Already a Backlash
OpenAI's update makes GPT-5 warmer and friendlier without increasing sycophancy, improving user satisfaction after feedback on its initial cold tone, with rollout starting August 15.
- OpenAI began rolling out a warmer, friendlier tone for GPT-5 on August 15, across web and mobile apps.
- A few weeks ago, users on Hacker News flagged GPT-5's tone, and OpenAI CEO Sam Altman admitted the rollout was 'a little more bumpy than we’d hoped for' after mixed reactions emerged.
- OpenAI says users will notice small, genuine touches like 'Good question' or 'Great start,' not flattery, and internal tests show no rise in sycophancy.
- According to help articles, users can shape responses by selecting tones like cheerful or cynical, and Plus and Pro subscribers can restore legacy models like GPT-4o via ChatGPT settings.
- So the real question is whether warmer GPT-5 will win back those who preferred GPT-4o, reflecting a broader trend in AI to balance personality and precision.
20 Articles
20 Articles
OpenAI’s long-awaited GPT-5 arrived this month with the kind of show that is normally reserved for the great technological revolutions. Sam Altman, its CEO and one of the most visible figures in Silicon Valley, had hinted that it could be a step towards the AGI—the so-called “general artificial intelligence” that would supposedly rival human thinking. But what we saw was another thing: a launch marked by overconfidence, embarrassing errors and t…
OpenAI says GPT-5 should be more friendly now
GPT-5 hasn't even been out for two weeks and OpenAI has already had to tinker with its tone. This Tweet is currently unavailable. It might be loading or has been removed. In a post on X over the weekend, OpenAI confirmed that it was making GPT-5 "warmer and friendlier" after initial complaints that it was too stodgy and cold in its interactions with users. Apparently, users should keep an eye out for phrases like "good question" in response to p…
After the release of the new GPT-5, many people indicated that they liked the previous version better.
Coverage Details
Bias Distribution
- 50% of the sources lean Left
Factuality
To view factuality data please Upgrade to Premium