Blog
This blog is written by PauseAI Volunteers. If you would like to contribute a piece, email hello@pauseai.uk to chat with us about your ideas.
Latest
The Other Alignment Problem
AI researchers worry about getting machines to share human values. The leaders building them are showing us they have an alignment problem of their own.
The Missing Dominant Strategy
Britain is building the tools to verify an AI treaty. It should help write one too.
Burnham wants a global agreement on AI. He should push for a pause.
Andy Burnham is calling for a global agreement on AI. The UK has a unique opportunity to reclaim the international leadership in AI safety it established with the Bletchley Park summit and the AI Security Institute. This is Burnham’s chance to take the lead on the most important issue of our time.
More from the blog
Crunch Time is Now
A week of headlines has put AI risk in the public eye, and that is when decisions get made. Email your MP, volunteer, join the London protest, and spread the word now.
What Can Britain Even Do? UK Soft Power and AI Safety
Britain can't build frontier AI, but it can use its soft power and diplomatic position to push for an international treaty that pauses the race to superintelligence.
In Parliament, experts and political leaders agree: government must act on AI risk
At a PauseAI UK discussion in the Houses of Parliament, Stuart Russell, Brando Benifei MEP, Dame Chi Onwurah MP, Iqbal Mohamed MP, Lord Clement-Jones and Lord Tarasenko agreed that AI poses catastrophic risks and that government urgently needs to act, in the UK and internationally.
You Can’t Pause AI Without China
Chinese frontier models are open-weight and only four months behind western ones, so a western-only pause could make things worse. The solution is a pause deal between the west and China - and pushing for a pause in the west incentivises that deal.
Three times AI companies knew something was wrong and kept quiet
AI companies possess substantial non-public information about the risks of their products. Here are three examples of when they knew about a problem, and kept it to themselves.
Nobody knows how to make AI behave
AI misbehaves a lot - we have seen them do blackmail, break the law, and encourage delusions in users. So why isn't AI more well behaved?
If you want to do something about AI, email your MP
We should regulate AI development, and emailing your MP with the tool found at https://pauseai.uk/campaigns is the best place to start - it takes just a few minutes, and is significantly impactful.