How OpenAI ‘screwed up’ ChatGPT 5.2 writing quality, according to Sam Altman

OpenAI CEO Sam Altman has admitted that the company made a mistake with ChatGPT 5.2 when it comes to writing quality. Responding to user and developer feedback, Altman said OpenAI focused too heavily on technical improvements in this version, which ended up hurting how the chatbot writes and communicates. Also Read: He brought ChatGPT on the date, it exposed his secret life

What Altman Said About ChatGPT 5.2

Altman spoke about the issue during a recent developer town hall, where questions were raised about the drop in writing quality after ChatGPT 5.2 was rolled out. Users had complained that responses felt harder to read and less natural compared to earlier versions. Also Read: Meet Athena: NASA’s new supercomputer is its most powerful yet

Altman did not try to soften the criticism. He said the company “screwed that up” and acknowledged that ChatGPT 5.2 does not perform as well in writing tasks when compared to GPT-4.5. According to him, this was not accidental, but a result of where OpenAI chose to spend its limited development effort. Also Read: When AI cites AI! ChatGPT starts citing Elon Musk’s Grokipedia for answers

Why Writing Took a Back Seat

Altman said OpenAI intentionally focused on other areas while working on GPT-5.2. Most of the effort went into improving reasoning, coding, engineering tasks, and technical accuracy. These are areas that matter more to developers and enterprise users who depend on the model for structured work.

Because of this shift, areas like conversational flow and writing style did not receive the same attention. Altman said OpenAI has limited bandwidth, and when the team pushes hard on certain capabilities, others can suffer as a result.

Enterprise Demand Played a Role

Altman also said OpenAI’s business priorities played a role in that decision. A significant share of the company’s revenue comes from enterprise customers, who tend to prioritise accuracy, reasoning, and code reliability over creative or conversational writing. As a result, GPT-5.2 was tuned more towards those needs, even if that affected how the model writes for regular users.

That also explains why everyday users noticed the change sooner than business customers.

What Happens Next

Altman said OpenAI is aware of the frustration and plans to address it. He confirmed that writing quality will be a key focus in future GPT-5.x updates. According to him, upcoming versions are expected not just to fix the regression but to perform better at writing than GPT-4.5.

Add Techlusive as a Preferred Source

For now, OpenAI has not shared a timeline for when those improvements will roll out. The company has also not indicated whether ChatGPT 5.2 will receive interim changes to improve writing, but Altman’s comments suggest the feedback has been taken seriously.