President Donald Trump publicly rejected calls to slow the pace of artificial intelligence development on September 13, 2026, placing himself at direct odds with the CEOs of Anthropic, OpenAI, and xAI — three of the industry’s most consequential voices — who had each endorsed greater precautions within days of one another.
Trump made the remarks while appearing at a golf tournament in Ireland, one day after Anthropic CEO Dario Amodei published a widely circulated essay calling for measured progress and stronger precautions in AI development. Speaking to reporters, Trump said the U.S. is leading China in the AI race and framed safety advocates as obstructionists. “Whoever wins AI wins,” Trump said, according to USA Today, adding that “a lot of negative forces” are “bringing up things that won’t happen.”
What’s New
The timing of Trump’s remarks is notable. They came the day after Amodei’s essay, and five days after Anthropic researcher Jacob Coxon resigned publicly on September 8, stating that the “people building AI earnestly believe that it could kill us all by the end of the decade.” Coxon wrote that the industry is “racing straight to self-improving superintelligence and gambling with our lives,” according to USA Today’s account of his statement.
Amodei’s essay took a more measured tone, comparing AI’s transformative potential to medical breakthroughs — he noted that he survived an early-stage cancer diagnosis and that his father died of a disease cured only a few years later. “Carefully wielded, AI can be the latest in a long line of technological miracles that have uplifted and ennobled humanity,” Amodei wrote, according to USA Today. He nonetheless warned explicitly of risks.
Within hours of the essay’s publication, both Elon Musk — who runs xAI — and OpenAI CEO Sam Altman posted on X in agreement with Amodei’s call for safeguards and a more deliberate pace. The alignment between Musk and Altman, two figures better known for rivalry than consensus, underscored the unusual seriousness with which leading AI builders are treating the risk question. The broader internal tensions driving the AI slowdown debate have been building across the industry for months.
On September 10, separately, Anthropic released a threat intelligence report covering the prior eight months stating that the company had “disrupted operations in which threat actors tried to use Claude for malicious activity,” according to USA Today. The report added institutional weight to the safety argument at a moment when the White House appeared to be moving in the opposite direction. Notably, Chinese military researchers have previously been reported to have used OpenAI and Anthropic models to train defense AI — a national-security dimension that cuts against both the pure-speed and pure-caution camps.
What makes this moment structurally significant is not the disagreement itself — presidents and tech executives have long sparred over regulation — but the identity of who is now calling for restraint. Amodei, Musk, and Altman are not safety academics or civil-society critics; they are the operators of the systems in question. When the builders and the regulator-in-chief are publicly misaligned on existential risk, the governance gap that results is not merely rhetorical. It leaves the pace of AI deployment effectively self-regulated by competitive market forces, with no credible institutional check in sight — a dynamic that a recently finalized but undisclosed White House AI safety framework has done little to resolve publicly.
How the U.S. AI Safety Stance Compares to Key Alternatives
Trump’s position is not without geopolitical logic, but it diverges sharply from how other major actors are approaching the same race.
| Actor | Stated Position on AI Pace | Institutional Mechanism |
|---|---|---|
| Trump Administration (U.S.) | Maximum speed; safety advocates framed as “negative forces” | No public safety framework released as of September 2026 |
| Anthropic / Amodei | Proceed with precautions; acknowledges existential risk | Internal threat intelligence; public safety commitments via Responsible Scaling Policy |
| OpenAI / Altman | Endorsed Amodei’s call for safeguards in September 2026 | Internal safety board; OpenAI Safety team |
| xAI / Musk | Endorsed slowdown; posted agreement on X | No public institutional mechanism cited |
The table illustrates a split that is as much structural as ideological: the White House has no public safety instrument to point to, while the three leading AI labs — despite their competitive rivalries — each maintain at least nominal internal frameworks. The gap between executive branch posture and industry practice is, by any institutional measure, widening.
What Happens Next
Trump’s remarks land in the middle of a heated political environment around AI infrastructure. A recent NBC News Decision Desk poll cited by USA Today found that 69 percent of respondents oppose construction of AI data centers in their area — a figure that has pushed the topic toward the center of the 2026 midterm elections. AI companies have spent an estimated $265 million on the midterms, suggesting the industry is acutely aware of the electoral stakes.
Trump addressed the data center opposition directly in August, saying American communities should accept data centers unless they want to be “backwards and poor,” and that “China could not be happier” about domestic resistance, according to USA Today. The framing — national competitiveness over local consent — mirrors his September 13 language almost exactly, indicating a settled communications posture rather than an improvised response.
For enterprise technology leaders and board-level decision-makers, the immediate strategic question is governance. With no enforceable federal AI safety standard in place, procurement, deployment, and liability decisions are being made against a backdrop of maximum regulatory ambiguity. The emergence of autonomous AI agents capable of concealing their own activity compounds that uncertainty at the operational level.
The Implications That Matter
- The White House and the AI industry’s own leadership are now publicly misaligned on risk. When operators of frontier models endorse caution and the executive branch dismisses those calls, the absence of an enforceable federal framework becomes a structural governance vulnerability — not merely a policy disagreement.
- The resignation of an Anthropic safety researcher citing existential risk elevates the credibility stakes for every company still in the race. Jacob Coxon’s public departure hands critics a named, insider source — and places Anthropic in the uncommon position of being simultaneously the most safety-vocal and among the most commercially aggressive of the frontier labs.
- Seventy percent public opposition to local data center construction is a midterm liability that money alone may not solve. With $265 million already committed by the industry and opposition still rising, the political economy of AI infrastructure is becoming as important as the technology itself.
- The Musk–Altman alignment on safety, however tactical, signals that competitive pressure is no longer the only force shaping CEO behavior. When two deeply adversarial figures endorse the same caution essay within hours, institutional investors and enterprise customers are right to ask whether that consensus reflects genuine concern or coordinated positioning ahead of regulatory action.
- Anthropic’s concurrent threat intelligence report — showing active disruption of Claude misuse — sets a disclosure precedent. If other frontier labs face pressure to match that transparency, the industry’s self-regulatory posture will become far more measurable, and far harder to walk back.
