Labour conference to vote on proposal for a new UK AI regulator

In short: Labour members will vote next week on a union proposal to replace the UK’s AI Security Institute with a stronger regulator.

What happened

Labour’s annual party conference is set to vote on a motion that calls for a new, more powerful regulator for artificial intelligence, also known as AI (computer systems that can produce text, images, or decisions that can look human-made). The motion was brought by the Communication Workers Union (CWU). The vote is non-binding, which means it does not force the government to act.

The CWU wants to replace the AI Security Institute (AISI), a government body set up three years ago that tests some of the most advanced AI systems. The union says AISI is under-resourced and has no real power because companies can choose whether to take part. A CWU official described the current setup as “opt-in,” like a safety check that firms can skip.

Government concern about AI is rising. Louise Haigh, the UK first secretary of state, told the Trades Union Congress that AI poses “huge risks” to national security without proper protections. Another senior minister told the Financial Times that AI is being discussed more often in cabinet meetings.

The debate has also been fuelled by recent disputes about AI testing. The Financial Times previously reported that Anthropic declined to submit its latest model to AISI for testing before release.

Why it matters

AI is spreading quickly into workplaces and public services, and people worry about job losses, surveillance at work, and safety risks. A stronger regulator could work like a tougher building inspector, setting clearer rules and having the power to enforce them, rather than relying on voluntary cooperation.

Source: Financial Times