UK Prime Minister Rishi Sunak hosted over 25 world leaders for a two-day AI safety summit this week, focussed on AI dangers and opportunities.
The world is pouring cash into AI right now, with the annual rate projected to hit $200B by 2025. And it’s impacting every sector, while also empowering citizens to brew better beer or add mullets to world leaders.
But the pace of change has instilled a sense of urgency around AI regulation.
So here are some quick highlights from the UK’s two-day summit:
- ✍️ 29 nations (including the US and China)signed the Bletchley Declaration on basic AI principles (eg, AI “poses significant risks”)
- 💻 Leading AI companies (Open AI, Google, Microsoft) agreed to work with governments to test their new AI models, and
- 📅 South Korea and France agreed to keep it all rolling, hosting the next two AI summits over the coming year.
Intrigue’s take: While some argue for the UN to lead here, it’s hard to see an effective UN process emerging in time, let alone keeping pace thereafter.
So in the meantime, states will keep racing to develop and deploy AI. Yet very few states actually have the resources or technological know-how to compete, and virtually all that do were gathered at Bletchley this week.
So to the extent we get meaningful AI regulation any time soon, our sense is it’ll look more plurilateral (like this week) rather than multilateral (via the UN).
And that’s probably better than nothing.