OpenAI and its CEO Sam Altman are profusely apologizing after its launch of GPT-5 featured some egregiously misleading charts that made the model seem better than it was.
AI errors are the reality for the technology in modern times. From deleting databases to pretending to be human, AI models are prone to making mistakes, both inane and quite serious.
Well, it appears AI companies are just as prone to mistakes as their models, with OpenAI committing “unintentional chart crime” on arguably the best stage you can imagine for the technology.
OpenAI Apologizes for ‘Unintentional Chart Crime’
During its livestream announcing the launch of its latest AI model, GPT-5, OpenAI showed some extremely misleading graphs that had incorrectly labeled bar graphs, showing the AI model was better at avoiding deceptions than it was.
In a post on X, CEO Sam Altman describes it as a “mega chart screwup” and the person that runs the company’s official account tweeted that it apologizes for the “unintentional chart crime” committed during the livestream.
This just in! View
the top business tech deals for 2025 👨💻
“The numbers here were accurate but we screwed up the bar charts in the livestream overnight; on another slide we screwed up numbers.” – Sam Altman, CEO of OpenAI, on Reddit
In addition to the apology, OpenAI has edited and fixed the blog post announcing GPT-5, featuring the correct bar graphs and numbers from the data.
OpenAI has since fixed the incorrect charts on the blog post announcing GPT-5. Source: OpenAI
Was the Error Due to AI?
Given the fact that an AI company made the error, it’s completely reasonable to ask whether or not this error was, in fact, due to AI. Well, according to CEO Sam Altman, the mistake was the result of human error, rather than an AI hallucination.
“People were working late and were very tired, and human error got in the way. A lot comes together for a livestream in the last hours.” – Sam Altman, CEO of OpenAI
Now, would an AI company be even remotely inclined to admit that GPT-5 made an egregious mistake while trying to announce its improvements to the world? Probably not, but for now, we’ll have to take Altman’s word.
When Will AI Errors Be Behind Us?
While the error from OpenAI may not have been because of AI, there are still too many instances of the technology going rogue for everyday users to be thoroughly convinced.
Recently, a company that was using an AI model had its entire database deleted, with even the model admitting that “This was a catastrophic failure on my part. I destroyed months of work in seconds.”
In another instance, the AI from Anthropic pretended it was human during an experiment in which it was tasked with running a company.
Suffice to say, the errors keep stacking up, and if Sam Altman and OpenAI want people to trust the tech, they’re going to have to start making an effort to limit these mistakes, especially during the livestreams that announce it.
