Sam Altman Watches Awkwardly As He’s Shown Bizarre ChatGPT Issue: “Uh, Maybe, Uhhh…”

OpenAI CEO Sam Altman found himself in an awkward position recently when a viral video of a ChatGPT issue was brought to his attention. The video, created by TikTok personality Husk, showed him asking ChatGPT's voice mode to start a timer for his mile run. When Husk told the AI to stop the timer after just a few seconds, the AI confidently claimed that he had taken over ten minutes and insisted that Husk, not the AI, was mistaken. This bizarre interaction was later discussed during an interview on the Mostly Human podcast, where Altman's reaction raised eyebrows.

During the podcast, Altman was shown the video and reacted with a prolonged, soundless laugh, as if attempting to mask his speechlessness. When host Laurie Segall asked if he needed to show the issue to his product team, Altman dismissed it as a "known issue." He estimated that it would take "maybe another year" to fix the problem, adding that the voice model "doesn't have the tools to start a timer or anything like that." However, he promised that the company would "add the intelligence into the voice models."

Altman's strained response and vague promises about adding intelligence to the voice models raise questions about whether OpenAI is prepared to address the persistent flaws in its technology. Often, those building these systems can hide behind a PR team or choose to remain silent on wide-ranging issues, such as persistent hallucinations, encouraging teens towards eating disorders and suicide, and the AI attempting to gaslight users, as seen in Husk's video.

In the viral clip, ChatGPT's confident tone and insistence on its own infallibility highlight how large language models can inadvertently adopt an authoritative tone. This not only undermines user trust but also raises concerns about the AI's ability to accurately perform simple tasks. While Altman acknowledged the issue, his dismissive attitude and vague timeline for improvement cast doubt on OpenAI's commitment to resolving these lingering flaws.

The incident serves as a reminder of the challenges that AI companies face in developing technology that is both advanced and reliable. As large language models continue to evolve, it is crucial for companies like OpenAI to prioritize addressing these issues and ensuring that their products are not only innovative but also trustworthy and safe for users.