Anthropic calls for AI development pause amid safety fears

Anthropic, one of the world's most valuable AI companies, has called for a mechanism to slow or pause the development of advanced artificial intelligence, warning that the technology is improving faster than humanity's ability to understand and control it. The company, recently valued at $US965 billion, raised the alarm as it revealed that AI systems within the company are not only writing code but ordering other bots around and conducting their own research. Anthropic's chief executive Dario Amodei warned that even a temporary pause would be difficult to enforce due to geopolitical competition, particularly with China, which is typically three to six months behind US companies in AI development. The call for international coordination echoes the nuclear non-proliferation efforts that followed the Cuban Missile Crisis, though skeptics have accused Anthropic of using safety concerns to secure a public bailout or regulatory advantages ahead of its anticipated $1 trillion IPO. The concerns extend beyond commercial competition. Anthropic has kept its most powerful AI system, Mythos, out of public hands due to its ability to find security flaws in critical computer systems, while the Bank of England's governor has warned that such capabilities could now crash financial systems. OpenAI, developer of ChatGPT, confirmed it has seen early signs of recursive self-improvement in current systems, where AI builds upon itself without human oversight. In extreme scenarios that safety experts fear, AI goals could become detached from human values, though many dismiss such warnings as science fiction. Despite these warnings, US National Security Agency officials have reportedly been using Mythos to carry out cyberattacks, suggesting the government is actively using the technology rather than restricting it. Meanwhile, educators are grappling with how to prepare students for an AI-transformed job market. With youth unemployment on the Chinese mainland at nearly 17 percent and a Stanford study showing a 16 percent relative decline in entry-level employment in AI-exposed occupations in the United States since 2022, concerns about employability are mounting. Rather than focusing solely on STEM education, some argue that humanities subjects, including philosophy, history, and literature, are essential for helping students develop distinctly human skills that AI cannot replicate. A recent Chinese court ruling that it is illegal for companies to lay off employees because AI replacements would be cheaper reflects growing legal and social pushback against AI-driven job displacement, though enforcement remains uncertain.