Britain Wants Real Rules for AI — OpenAI Surprisingly Agrees

Britain has spent years trying to regulate artificial intelligence with a light touch: existing watchdogs plus voluntary promises from AI companies. Now Parliament's Joint Committee on Human Rights says that approach isn't working and wants a full AI law, with an independent regulator, mandatory transparency, and powers to check what these systems actually do. Committee chair Alex Sobel MP said no country, including the UK, has rules that are "fit for purpose." Separately, more than 70 MPs and peers backed calls to block development of artificial superintelligence — AI smarter than humans, which doesn't exist yet.

The surprise is who agrees: OpenAI. The company behind ChatGPT now says governments should impose mandatory safety requirements on the biggest AI labs, including independent testing. That's partly self-interest. Expensive compliance is easier to afford when you have billions in the bank, and strict rules can squeeze out smaller rivals. But OpenAI's core argument holds up: companies shouldn't be the only judges of whether their own products are safe. We don't let drugmakers quietly decide whether their medicines have been tested enough.

Why should you care? AI is already in your life: chatbots answering your customer queries, software screening job applications, tools summarising your medical notes. Today, when something goes wrong, there's often no legal body to complain to and no requirement anyone told you AI was involved. A real law would mean public safety tests before powerful models launch, and someone accountable when they cause harm.

The catch: this is a proposal, not law, and "regulation" can mean anything from meaningful safety checks to paperwork that protects big incumbents. Superintelligent AI also doesn't exist yet, so some of the loudest warnings are about a hypothetical future rather than your inbox today.