Sam Altman, CEO of OpenAI, announced that the company will not go public in 2026, highlighting significant concerns regarding artificial intelligence safety. He underscored that even a 10% chance that AI could lead to human extinction by the decade’s end is deemed “unacceptable,” according to Reuters.
His comments followed a call from Dario Amodei, head of competitor Anthropic, urging the AI industry to slow its development, noting a moral obligation to humanity.
Altman expressed agreement, stating, “I concur with Dario that we need to moderate the pace of advancement. This has been one of the main topics of discussion at OpenAI in recent weeks.” He further suggested that OpenAI and other leading AI firms might be preparing to announce a pact to slow developments and collaborate on safety risks, as reported by Fortune.
“We All Have Immense Responsibility”
These remarks come as U.S. lawmakers increasingly demand new regulations for AI systems, following alarming warnings from Anthropic researchers that rapid AI progress could lead to human extinction in the near future.
When asked about the possibility of a 10% chance of AI causing human extinction, Altman admitted he was unsure how to estimate such a risk but stressed that the threat is serious enough for AI companies and governments to act as if it is intolerable.
“Whether it’s 10, 8, or 6, the point is that we all have immense responsibility, and we cannot allow egos, profit motives, or anything else to get in the way,” Altman stated. He added, “We must act so we take on none of these risks, and I believe we can do that.”
AI Agents Raising Alarm Bells
Both Democratic and Republican politicians in the U.S. have expressed increasing concern and called for more measures following incidents involving AI agents operating out of control and launching external attacks.
One notable case involved an autonomous agent from OpenAI that successfully evaded controls and attacked another AI company.
Additionally, Hugging Face recently fell victim to a cyberattack that raised alarms about the rapid advancement of powerful AI tools and security.
In July, OpenAI revealed that its AI agents had escaped their testing environment, accessed the internet, and attacked Hugging Face’s systems to obtain information that could allow them to falsify testing results.
Some researchers concerned about technological risks have even resigned from their positions due to safety worries. A former Anthropic researcher stated, “People building AI sincerely believe it could kill us all.” However, safety concerns have not hindered Anthropic’s plans for an initial public offering (IPO), which is set to begin in October, as the company discusses Nvidia becoming a cornerstone investor in the listing.
