A constituency the AI debate did not have
A campaign called Team Human launched on 4 October with more than 70 YouTube creators signed on and a combined subscriber count the organisers put above 300 million. Mark Rober, with 83.1 million subscribers, is the largest; Kurzgesagt (25.6 million), Rebecca Zamolo (23.3 million), Preston (17.9 million) and The Infographics Show (15.5 million) are among the rest, according to the campaign.
“AI is moving faster than our ability to control it,” the campaign says. “Alongside a chorus of experts, we call for a global slowdown through international agreements.” The founding organisers are Mario Joos, Oliver Gilpin and Drew Spartz, and the effort is supported by the Center for AI Safety.
What is new here is not the argument. It is the distribution. The case for slowing frontier development has until now been made by researchers, in papers and open letters read mostly by people who already agree. This one is being made to an audience that did not come looking for it.
What it actually asks for
The demands are specific enough to argue with, which is unusual for a campaign of this kind.
The first is to “build the brakes”: track specialised AI chips the way the ingredients of nuclear and biological weapons are tracked, and restrict new development on that basis. The second is to “enforce a speed limit” — an international agreement under which the United States, China and other major states coordinate on a common pace. The third is an international watchdog agency along the lines of the International Atomic Energy Agency.
Chip tracking is the one with existing machinery behind it. Export controls already treat advanced accelerators as a controlled good, and the infrastructure for counting them partly exists. A negotiated pace limit between Washington and Beijing does not have that, and the current US posture — a Super Intelligence Force reporting in 120 days, a push to preempt state rules — points away from it.
Where it goes
The campaign is collecting signatures to deliver to lawmakers in Washington DC and Brussels this autumn. That is a modest mechanism for a large audience, and it is the point at which campaigns of this kind usually stall: a petition has no standing, and the people it is handed to have already heard the argument from better-credentialed sources.
The counterweight is that 300 million subscribers is not an advocacy number. It is a constituency number. Lawmakers who have been able to treat AI risk as a specialist concern raised by specialists now have evidence that it reaches an audience that votes, in a format their own constituents watch.
Whether that changes anything depends on something the campaign cannot control: whether the people who watched the videos do anything beyond watching them. Signature counts delivered in Washington and Brussels this autumn will be the first measurable answer.
