Wednesday, September 23 September 23, 2026

Good morning. It's Tuesday, September 23rd, 2026. Welcome to the MorningAI daily briefing. Today, three very different visions of AI's future collided on the world stage — and the tension between them is going to shape the next decade of this industry.

Let's start at the United Nations General Assembly, where AI dominated the agenda.

President Trump took the stage and made a name change official. The United States, he declared, will no longer call it artificial intelligence. Going forward, the administration is calling it "Super Intelligence." That's not just branding. Trump used the move to double down on his anti-regulation stance, telling world leaders that the U.S. Justice Department could rein in AI companies if needed, but that his administration has no intention of letting a quote "globalist scheme" dictate how American AI develops. The message to Silicon Valley: you're free to build, just don't embarrass us.

Meanwhile, the United Kingdom went the other direction. The UK announced a new global AI safety standards initiative at UNGA, paired with a coordinated campaign against AI-driven disinformation. The UK is positioning itself as the adult in the room on AI governance — a role it's been cultivating since hosting the first AI Safety Summit at Bletchley Park. This sets up a meaningful transatlantic split on how to handle the most powerful technology humanity has ever built.

And then there's Bernie Sanders. Senator Sanders and Representative Greg Casar dropped a bill this morning that would ban artificial superintelligence outright and create a new federal Department of Artificial Intelligence. This isn't a fringe position anymore — it's a legislative marker. The bill almost certainly won't pass, but it shifts the Overton window and gives progressive critics a concrete rallying point as AI capabilities accelerate.

On the model front, the Grok versus Claude rivalry continues to heat up. xAI's Grok 4.7 launched with aggressive pricing — two dollars per million input tokens, six dollars per million output — but independent benchmarks show it trailing Claude by roughly seven points across key evaluations. Meanwhile, Claude Opus 5.5 details are circulating across the AI research community, with early comparisons putting it against GPT-6 Astra and other frontier models. The race is very much alive, and pricing pressure is real — which is good news for developers and bad news for anyone betting on a single model winner.

On the regulatory and public sentiment front, a new NPR report highlights a growing debate around an AI slowdown — and the irony is striking. The loudest voices calling for a pause include some of the biggest AI labs themselves. Critics point out that an AI freeze would entrench incumbents and lock out smaller players and startups. If you're building, this is a moment to build faster, not slower.

And finally, Bill Gates published a major essay this morning on what he calls "the turbulent AI era." Gates argues the choices made right now — on safety, access, and governance — are more consequential than any individual product launch. It's a rare moment when the founder of the PC era is the voice urging caution while the next generation presses the accelerator.

The through line today is clear: the world is choosing sides on AI governance in real time, and those choices will determine who gets to define what Super Intelligence — or artificial intelligence, depending on your politics — actually becomes.

Here's today's business idea: with the UK and US taking opposite positions on AI standards, there's a clear opening for a compliance intelligence platform that tracks global AI regulatory divergence in real time — built for multinational companies, law firms, and policy teams who need to know, right now, which jurisdictions require what. Think Bloomberg Terminal meets AI policy. The timing is perfect, the pain is real, and there's no dominant player in the space yet.

That's your MorningAI briefing for September 23rd. Stay sharp out there.