A White House meeting with AI executives ended with an agreement favouring industry self-regulation over government intervention to slow new models, El País reports. The account captures a pointed split over who should manage AI’s risks: Nvidia’s Jensen Huang argues that companies should solve safety problems through engineering, while Anthropic’s Dario Amodei is identified as a leading advocate of slowing development.
Watch Desk analysis
What happened
El País describes a meeting between Donald Trump and senior AI industry figures, including Google’s Sundar Pichai, Meta’s Mark Zuckerberg, Elon Musk, Huang and Amodei. The paper says the firms signed an agreement calling for self-regulation and rejecting government steps to slow the progress of new models.
The article also recounts Huang’s argument that companies should not ship systems they cannot make safe. That is his position as presented by El País, not evidence that the industry has settled how safety should be measured or enforced.
Why it matters
The disagreement is about more than executive nerves in front of cameras. It goes to the practical question of who sets the limits for increasingly capable AI: companies building the systems, or governments with power to regulate them.
The reported agreement points towards the first answer, while Huang’s “don’t ship it” argument puts the burden on companies to treat safety as an engineering problem. What remains unclear from this account is how the agreement would work in practice, or what oversight and consequences it includes.
Our read
The contrast is worth watching: self-regulation sounds less reassuring when the people being asked to regulate themselves disagree about how much danger there is. Huang’s engineering framing is a useful challenge to vague alarm, but “solve it before shipping” still leaves the important questions of standards, proof and accountability.
What to watch
- Whether the agreement leads to specific safety standards or remains a statement of principle.
- What role, if any, government oversight retains.
- Whether AI companies can show how their safety claims are tested and enforced.
Discussion spark: Should AI companies be trusted to set and enforce their own safety limits, or should governments have the power to slow releases when risks remain unresolved?
Sources and evidence
- AI’s top leaders, like children afraid of a bully – EL PAÍS English (4 October 2026, 04:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.