White House AI Safety Pledge Meets a Hard Truth: Only 2% of Consumers Are Sold
Tech CEOs signed a "morally binding" safety pledge as new data shows consumer appetite for AI remains at roughly 2%, even as Washington rebrands the technology as "super intelligence."
The White House this week gathered nearly every major tech CEO in one room — Mark Zuckerberg, Jeff Bezos, Elon Musk and Anthropic's Dario Amodei among them — to sign a voluntary AI safety pledge that President Donald Trump described as "morally binding." Trump also signed an executive order that formally rebrands artificial intelligence as "super intelligence" across federal language. The symbolism was hard to miss: an industry under scrutiny standing shoulder to shoulder with the administration that regulates it.
The pledge carries no enforcement mechanism, and that is the point. Signatories commit to vague principles rather than auditable benchmarks, leaving compliance to the companies themselves. Anthropic's presence is notable given its safety-first positioning, while Meta and OpenAI have spent recent months softening their public messaging, rolling out friendlier product faces and consumer features that downplay existential risk. The same firms signing safety promises are competing to make AI feel harmless enough to use every day.
That marketing push runs into a stubborn problem. Consumer surveys cited in this week's coverage put actual AI adoption — people paying for or regularly relying on the tools — at roughly 2%. The gap between the executive enthusiasm on display in Washington and the number of ordinary users opening their wallets is the defining tension of the moment. Billions in infrastructure spending are being justified by demand that has not yet materialized at scale.
Washington's rebranding effort may widen that gap rather than close it. "Super intelligence" is a term built for researchers and investors, not for someone deciding whether an assistant is worth $20 a month. Renaming a technology through executive order does not change what it does for a user, and the phrase invites the kind of sci-fi anxiety that consumer-facing teams at Meta and OpenAI are working to avoid. Two audiences are being addressed with two contradictory stories.
What to watch next is whether the safety pledge produces anything checkable — an audit, a disclosure requirement, a third-party review — or joins the pile of voluntary commitments that fade after the photo op. The harder question is commercial: if adoption stays near 2%, the industry's spending plans and the administration's superintelligence framing will need a base of real users that does not currently exist. The CEOs have signed. The customers have not.