On Saturday, September 12, leaders of three of the biggest firms working on Artificial Intelligence seemed to agree that it was time to moderate the pace of development of the frontiers of AI. I take that to be a good thing — not enough, but a good thing.
It’s unclear if anything will actually happen differently after this weekend. These are, after all, just words from corporate actors with a long track record of mostly just charging ahead. And they are doing so in a field with very few meaningful limits from governments of any kind. So we will take these words of caution with a very big grain of salt for now.
And political leaders have expressed concerns about the harms associated with AI this weekend — though likewise, very little has actually changed. Bills are pending in legislatures across the 50 U.S. states, the U.S. Congress, and around the world. I’m told that a pile are actually on the desk of a single governor, Gov. Gavin Newsom of California, at this very moment.
The ideas that got some additional traction this weekend are all sound: (1) allowing genuine third-party verification of alignment mechanisms, (2) adopting some guardrails on technical development of AI tools, and (3) international cooperation, and perhaps treaties, to take away the risk of a competitor nation racing ahead and gaining a permanent geopolitical edge. In the blog post from Dario Amodei that got a lot of attention this weekend, he distinguished between two types of coordination — one between and among democratic societies, and the other among all nations. (On this last point (3), be sure to check out also the thoughts of Verity Harding and the AIxGeo initiative at the Bennett School at the University of Cambridge.)
As I spent a lot of time listening to the words a few of these “titans” of the big AI firms — Dario Amodei of Anthropic, Sam Altman of OpenAI — I have also been thinking: to whom else should the world be listening right now on matters of AI development? There are lots and lots of people in addition to these two who deserve the microphone.
To this end, I posted a note to LinkedIn to gather some good names of others to whom we should listen right now. I suggested one of my own (Dr. Alondra Nelson of the Institute of Advanced Studies, and a MacArthur Foundation Board member). Many others have chimed in with dozens more.
The good news: there are a lot of voices, from diverse backgrounds, lots of brilliant people who have been sounding the alarm (and cheering the successes) about AI for some time now. It’s high time we take their views into account — and do something about it.
For instance, have you taken a look at Urs Gasser and Viktor Mayer-Schönberger‘s book, Guardrails? Lots of ideas in there and in their other writings about what might be done.
Or Martha Minow‘s 2025 Rene Cassin lecture and essay at McGill?
Or the extensive work of Catherine Bracy and her team at TechEquity?
Or the new report from Data + Society on Governing AI agents published this week by Janet Haven and Rangit Singh?
Please add more names — here on this blog or on LinkedIn — and I’ll tune into them, too.
For now, the options are plentiful and the good ideas are out there. It’s not as though pure “industry self-regulation” or “do nothing” or “roll the dice” are the only answers. And it’s not too late — yet — to do something, or better yet, some things.
At MacArthur Foundation, we’ll keep working with our peers to keep funding and supporting them. I hope you’ll join us in giving the mic to the incredible people who should be heard in these debates and leaning into the power of the broad swath of humanity — not just the few, and not just those with a profit motive or a political bent towards only letting industry decide what is best for the other 8 billion or so of us.
Discover more from John Palfrey
Subscribe to get the latest posts sent to your email.