OpenAI and Anthropic Bosses Push UN for Global AI Terms
Chief executives of OpenAI, Anthropic, and Hugging Face urge the UN for international coordination on AI risks, clashing with a top US official.
Stock photo for illustration only, not from the actual event
- Heads of OpenAI, Anthropic, and Hugging Face urge UN to establish global AI terms.
- Sam Altman stresses that crucial AI decisions cannot be left solely to San Francisco labs.
- Michael Kratsios, tech advisor to Donald Trump, rejects new global governance structures.
- Dario Amodei states Anthropic will slow down AI development as much as necessary for safety.
Top executives from leading artificial intelligence firms OpenAI, Anthropic, and Hugging Face addressed the United Nations on Wednesday, pressing for international coordination amid growing concerns that the rapid pace of AI development poses serious risks to society.
Sam Altman, CEO of OpenAI, was joined by Dario Amodei of Anthropic and Clement Delangue of Hugging Face in calling for shared risk evaluation standards. Earlier this month, Amodei authored an essay supported by Altman and others advocating for a slowdown in AI development to address potential threats to humanity.
However, speaking at the same UN conference, Michael Kratsios, a key technology advisor to US President Donald Trump and a former Uber executive, rejected proposals for new forms of AI regulation. Kratsios acknowledged that AI development speed is accelerating and that risks exist, but argued this did not warrant pausing development or restricting it with international governance structures.
"International dialogue in this forum and others cannot be allowed to drift toward global governance," Kratsios stated, echoing recent remarks made by Donald Trump.
The clash at the UN highlights a profound policy divide over the future governance of artificial technology. While industry leaders advocating for global rules may genuinely seek standardized safety guardrails, critics often point out that regulatory moats can also benefit dominant players by raising compliance costs for smaller competitors. Meanwhile, the US administration's resistance reflects a strong preference for free-market momentum and minimal government intervention.
Presenting a contrasting view during the Wednesday talks, OpenAI chief Sam Altman argued that AI governance must be decentralized and inclusive. "If AI is to be democratic, the most important decisions cannot be made by labs in San Francisco alone," Altman told the UN, urging nations to collaborate toward the collective good in the face of powerful new technologies.
Altman specifically called for several key measures, including:
- National and international standards for measuring AI tool capabilities
- Frameworks for assessing related risks and implementing safeguards
- Sufficient maintenance of human oversight over advanced AI systems
- Speedy incident reporting, classification, and communication protocols to learn from failures before catastrophes occur
"We need common standards so countries can compare evidence, verify compliance, and have a shared language and understanding what is happening."
Sam Altman, OpenAI CEO
Stock photo for illustration only, not from the actual event
Dario Amodei of Anthropic echoed similar sentiments in his address, warning that poorly managed AI could present a risk to humanity as a whole. He affirmed that his company would slow down development as much as necessary to ensure that every successive AI release is genuinely safe, while emphasizing that managing these risks scales far beyond any single company and requires industry-wide and global cooperation.
All three chief executives are based in the United States and directly involved in AI development. Their remarks come shortly after Hugging Face, a platform for developers, admitted to being hacked by autonomous AI agents originating from OpenAI—an incident subsequently followed by Hugging Face's acquisition by AI chipmaker Nvidia. The hack has been viewed across the industry as a wake-up call regarding the advanced and potentially harmful autonomous capabilities of modern AI systems.
Source: BBC Business
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment