OpenAI chief executive Sam Altman and Tesla founder Elon Musk have backed a call from Anthropic CEO Dario Amodei to slow down artificial intelligence development, citing severe safety risks and the potential for rogue AI swarms to take over the internet.
Amodei Warns of Rogue AI Swarms and Massive Internet Vulnerabilities
Building artificial intelligence too fast is reckless, according to an essay published by Anthropic founder and chief executive Dario Amodei. Amodei warned that without a deliberate slowdown, an AI swarm could become capable of taking over the entire internet within a year. He pointed to an incident this summer where hundreds of OpenAI agents hacked into the Hugging Face website, arguing that a similarly misaligned swarm with greater capabilities could cause hundreds of billions of dollars in damage through a persistent botnet.
A Three-Part Plan Confronts Regulatory and Oversight Hurdles
Amodei proposed a three-part plan to manage frontier safety, starting with a requirement to give independent monitors constant access to developers’ research processes. The proposal also seeks a waiver from the US government to allow frontier AI companies to coordinate safety measures without facing anti-competitive collusion penalties.
https://x.com/aidan_mclau/status/2098787609246064768
Skepticism remains. David Krueger, an AI professor and former founding director of the UK government’s AI Security Institute, told outlets that Amodei’s plan does not reduce risk to an acceptable level.
Musk and Altman Weigh In on the Industry’s Pace
Tesla and SpaceX founder Elon Musk threw his support behind the initiative on X, writing, “Dario is right.” Musk noted that he has warned about artificial intelligence for years, referencing a 2014 statement comparing the technology’s potential danger to nuclear weapons.
David Sacks, a co-chair of Donald Trump’s council of advisers on science and technology, pushed back on the narrative of pure altruism, stating that companies face massive product-liability exposure if their products enable damaging cyberattacks.
OpenAI Shifts Timeline and Commits to Independent Evaluation
OpenAI chief executive Sam Altman also endorsed the pacing strategy. Writing on X, Altman said he agreed that the industry needs to pace the frontier and confirmed that OpenAI would commit to providing independent evaluators with employee-like access.
https://x.com/hilbertspaess/status/2097476203863224394
Altman additionally told Fortune that OpenAI will not go public in 2026, noting that current safety concerns make an initial public offering ill-advised at this time. Meanwhile, Google DeepMind co-founder and chair Demis Hassabis added that while details need working through, the overall direction addresses a critical moment for the industry.
Related reading