TL;DR

More than 100 UK parliamentarians across parties are demanding binding regulations on the most powerful AI systems. The cross-party group includes former ministers warning that superintelligent AI “would compromise national and global security.”

Growing Pressure for Regulatory Action

A coalition of Westminster MPs, peers, and members of devolved legislatures is calling on the government to move faster on AI regulation. The campaign, coordinated by nonprofit Control AI, urges Keir Starmer to show independence from the Trump administration’s opposition to AI regulation.

Former defence secretary Lord Des Browne describes superintelligent AI as “the most perilous technological development since we gained the ability to wage nuclear war.” Former environment minister Zac Goldsmith criticises governments for being “miles behind the AI companies” whilst leaving development “with virtually no regulation.”

Industry Warnings Add Weight

The calls coincide with warnings from within the AI industry itself. Anthropic’s chief scientist Jared Kaplan recently told The Guardian that humanity would need to decide by 2030 whether to take the “ultimate risk” of letting AI systems train themselves. One of AI’s pioneers, Yoshua Bengio, has noted that AI is currently “less regulated than a sandwich.”

Government Response

Labour’s 2024 programme committed to legislating on powerful AI models, but no bill has been published. The government faces reported White House pressure not to inhibit commercial AI development. A Department for Science spokesperson noted that “AI is already regulated in the UK” with existing rules in place.

Looking Forward

Control AI’s chief executive Andrea Miotti warns that mandatory standards could be needed within one to two years given the pace of AI advancement. The UK’s first AI minister Jonathan Berry suggests global rules should include “tripwires” requiring testing, off switches, and retraining capabilities for models reaching certain power thresholds.

Source: The Guardian