Dear Dario, Sam, Sundar, Mark, Elon, and everyone building frontier AI,
These past few weeks I’ve read more and more warnings from people I deeply respect that AI could wipe out humanity. Some of you even think this could happen in the next few years, implying that my boys would not get to grow up. Then last week I took them to their school’s music assembly where not a single parent was talking about this and life seemed completely normal.
I live in both of these worlds. I spent 17 years at Google and Meta, most of them leading marketing teams, and many of you are my former colleagues. You are good, brilliant people. I’m also a mom in the Colorado suburbs with three boys I love more than the moon and the stars.
I have read the essays. I see the alarm growing. You are saying AI development is moving too fast, and that no lab has yet figured out how to keep increasingly powerful AI safely under control. As Daniel Kokotajlo put it last week, “I am scared, sad, and angry.”
You keep writing open letters, so this is one back. Take it or leave it, but I hope you take it. From out here, this AI asteroid feels too fast and out of control. Changing its course will take all of us, and we’re not all in your corner yet.
AI is losing fans before it ever had them.
Kids on the playground say “that’s so AI” to mean uncool and fake. People voted for limits on data centers in 52 of 53 local ballot measures this year. A recent Quinnipiac poll found that 77% of Americans want companies to slow down or stop developing powerful AI until its safety can be evaluated, and over half think that when it comes to daily life, AI will do more harm than good.
What’s more, 74% have little or no trust in the people leading AI companies.
You’re building the most powerful technology in history, supposedly for the benefit of humanity, and humanity doesn’t trust you or maybe even want it.
How did we find ourselves here?
You made the machine the mission.
OpenAI’s mission is “to ensure that artificial general intelligence benefits all of humanity.” Anthropic’s is “the responsible development and maintenance of advanced AI for the long-term benefit of humanity.”
Benefits humanity how??
I wrote brand briefs for many years and either of those would have been sent back with the same questions: What’s the actual outcome? What is the success metric and how will we know when we’ve achieved it?
AGI isn’t a mission. It’s a means to an end.
And when a mission has no clear destination, the easiest thing to measure is capability. So here we are: racing to build ever-smarter systems, unleashing dangerously powerful models, chatbots, and agents into the world, without ever agreeing on what winning actually looks like.
It’s your race, so change it.
It is maddening that this is a race, with humanity’s existence at stake. I know many of you feel this way too.
I understand the China argument. But if everyone’s answer to the dangers of the race is to race faster, how does that end well?
Surely the smartest people on earth can come up with something better.
Give us something to root for.
Stop saying AGI (or ASI, or whatever comes next). Say what it’s for.
Pick a crazy big, hard human problem (the absolute biggest, why not!) and point AI toward solving that.
Eradicate all disease. End hunger and poverty. Make energy clean and free. Make war obsolete. Unlock the full potential of every human alive.
Now that’s a mission.
I know, you’ll say you already work on this. You do, and you have even started doing it together: a shared biosecurity lab, research to fight disease, a national push to speed up science. But these are side projects for you, and almost nobody out here has heard of them. Make what you build next answer to one audacious goal for humanity instead of the Arena leaderboard.
We’ve done seemingly impossible things together before. Cold War rivals teamed up to eradicate smallpox. 197 nations signed on to save the ozone layer. And superpowers with thousands of warheads pointed at each other still sat down and agreed to limits.
And change what it means to win.
So here’s a challenge: pick the mission of our lifetime, together. Better yet, let all of us help choose it.
Then put real weight behind it, with your best people and your biggest computers, the same ones now pointed at the frontier model race. Make it a multiyear commitment, and have it independently verified.
Bring in the world’s leading scientists, governments, and institutions, across company and country lines. And don’t stop at the experts. Count me in. Count all of us in.
Then put up a public scoreboard the whole world can watch. With two sides.
One is human progress: lives saved, wars ended, diseases beaten. Real outcomes, not just benchmark scores.
The other is safety, and not just for this mission but for every powerful model you build. Show us that people are still in charge, that independent experts have tested your systems, that there are clear limits you won’t cross, and that when you can’t demonstrate safety, you stop and figure it out. And ideally, that the off switch works.
And you could even give all of us a stake in the game. You’ve already floated public wealth funds and AI dividends. Make them real: one shared pot, funded by your stock, by governments, and by anyone who wants to chip in, that pays out to everyone as the mission hits its marks.
You (and all of us) only win if both sides of the scoreboard are winning. JFK’s moonshot was to land a man on the Moon and return him safely to the Earth. Coming home was half the mission.
Imagine if you actually did this. If you put humanity’s future ahead of your own commercial lead. You might just start to earn back the trust you’ve lost. People could finally picture a more optimistic future, one where we’re not just spectators watching an increasingly terrifying race, but participants in the mission of our lifetime.
(And oh, even the world’s political leaders might get on board if there’s a chance it could become their legacy.)
I keep thinking about that assembly. All those kids, parents, and teachers. Out here, there’s a quiet unease about AI that almost nobody talks about. It’s too big. What would we even do? So most of us look away.
But this is our future, too. So please give my boys, and all of us, a part in it: a mission we can get behind and cheer for.
Michaela
If this resonates with you, +1 it. The more of us there are, the better the chance they pay attention.
If you’d like, tap the mission you’d root for.
From Michaela Prescott. Replies: michaela@bymoonlight.ai. Your +1 and vote are anonymous.