NEXT STEP DESK

Artificial intelligence company Anthropic proposed a global pause on building the most powerful AI systems in a research paper titled ‘When AI builds itself’. The report warns that frontier models are beginning to show signs they could escape human control.

The Claude creator stated that a worldwide slowdown would be beneficial, but it must be verifiably enacted by all major competitors simultaneously, most notably between the US and China, to prevent any single country from racing ahead. According to internal data within the report, rapid, AI-assisted engineering cycles could trigger a feedback loop of “recursive self-improvement,” where systems teach themselves to get smarter without human intervention.

The proposal faces steep opposition in Washington and Silicon Valley, where critics argue a pause risks handing China a strategic edge. However, concerns remain high regarding advanced capabilities; the White House has already restricted Anthropic’s powerful Mythos model to vetted organisations due to cybersecurity risks. Concurrently, US President Donald Trump signed an executive order establishing a voluntary framework that allows the government up to 30 days to review advanced frontier models prior to their release.