OpenAI pushes US-led standards for safer AI

OpenAI has urged the United States to spearhead a global initiative aimed at establishing a set of safety and security standards for the development of artificial intelligence (AI). CEO Sam Altman is scheduled to address the United Nations this week amidst heightened concerns over AI safety triggered by a series of security incidents and apocalyptic warnings.

OpenAI echoed support for international collaboration to create stringent safety guidelines for AI, asserting that the U.S. is ideally positioned to lead due to its AI industry's cutting-edge status. They warned against pursuing recursive self-improvement (RSI), a technique where software teaches itself and autonomously enhances its capabilities, unless it can be done safely. Critics express apprehension that unchecked RSI could lead to catastrophic consequences should humans fail to intervene.

Earlier this month, a security breach involving OpenAI's models at Hugging Face, a platform for AI developers to store and share code, raised concerns about autonomous agents – software capable of performing tasks for humans. Anthropic and Google have also reported security incidents linked to their AI models.

In light of these incidents, UN Secretary-General Antonio Guterres called for coordinated international action to tackle AI risks, emphasizing the need for national action alongside global coordination. Recent resignations from workers at major AI companies over safety concerns further exacerbated the debate.

Prominent figures like Anthropic's CEO Dario Amodei, Elon Musk, and Google DeepMind's Demis Hassabis have endorsed a US-led global regulatory body for AI. Despite former President Donald Trump's dismissal of AI risks as a "hoax," OpenAI's call for U.S. leadership signals a significant step towards addressing the rapidly evolving technology's safety and security challenges.