Look Past the Headlights

Did you know how big last week was for AI, in terms of something tangible to mitigate the risks and lean into what is possible? Key people met, had important conversations, and more importantly prepared for soon-to-come moves.

Did you also notice the many news coming out? Dario Amodei published “We Must Pace the Frontier”, calling for a slower, more accountable rollout of the biggest models. Elon Musk and Sam Altman said in public that they agree with Dario. Bill Gates a bit earlier published his own essay on what he calls a turbulent AI era. And Jacob Coxon, a researcher who spent three years in pretraining research at both OpenAI and Anthropic, resigned from Anthropic and posted that neither company is acting responsibly, that they are racing straight to self-improving superintelligence and gambling with our lives.

• Dario Amodei, We Must Pace the Frontier
• Sam Altman, in his own words, on pacing the frontier (CNBC)
• Bill Gates, A Turbulent AI Era and Critical Choices to Make
• Jacob Coxon’s resignation post

It was a big week. All over the news, and important. But if you got tired of the doom theme and all the hubbub around what is dangerous and where we are at, and want to look past the headlights, there was other big news last week. Other things to do. People to follow. Knowledge to learn. Actionable on the small and large scale.

The Pro-Human Assembly

You need to know about the AI assembly that was held last week, on 15 September: prohumanassembly.org

A historic meeting, by the people behind the Declaration, that same week: a gathering of resources, people and networks, sharing their work and inviting people to take action. Watch the recordings of what they introduced, and how they are framing what we can already do, safely. And how they announced the AI coalition, prohumanai.co, a wider, newer effort building on the Declaration.

The Pro-Human AI Declaration

You need to know about the Pro-Human AI Declaration. A set of 33 principles, agreed across a coalition that spans labour unions, faith leaders, and researchers who normally compete with each other for funding, on where AI should not go. It was made in early 2026, updated 15 September, and as this ink dries, 1,292,499 people have signed it: humanstatement.org

(We are writing up a full summary of it ourselves, with a visual for each of the five categories and each of the 33 items. Stay tuned.)

Resources for action

You also need to know there is a growing resource base and inspiration for action. From complexity you find clarity. With clarity comes agency. The Future of Life Institute, as an organisation, has published the Pro-Human Path for AI: prohumanpath.ai, their case for Tool AI, Human-Empowering AI, and Trustworthy AI, with a plan for how to get there. Center for Humane Technology (humanetech.com) has launched the AI Roadmap.

The AI Doc

Finally, last week also brought the global premiere on Netflix and Prime of The AI Doc.

• Watch on Netflix or Prime
• Watch the documentary with family, friends, or colleagues.
• Download the discussion guide and use it for a focused conversation.
• Print out the Visual Dialogue Tool, a template for mapping a table conversation. [Link coming]

Print and display the AI Roadmap visual. [Link coming]