By Murray Hunter
INVESTORS are calling for AI development to be closely monitored and slowed down.
They are demanding that the state regulate AI.
The public is being sold a story about the dangers of AI to create a fear factor. Meanwhile, China is undercutting the cost of using Anthropic by a factor of a hundred.
The fate of the AI surge is in the balance.
Even Bill Gates has joined the chorus.
He proposes taxing AI tokens to slow the technology’s advance, framing it as concern for human jobs that require “empathy”, education, and care.
In the space of a few days, Anthropic founder Dario Amodei publicly urged external evaluators to receive the same access as company employees.
Sam Altman of OpenAI and Elon Musk of xAI all immediately concurred.
A researcher resigned from Anthropic, declaring the race for superintelligence an existential threat.
Netflix released a documentary weighing AI’s benefits against its dangers.
Klaus Schwab’s circle dusted off “Cyber Polygon” language, the familiar cascade-of-crises framing that once accompanied pandemic planning.
The media is amplifying every note, quickly cashing in on a new narrative.
Headlines are warning that AI could kill all humans within a decade
The same outlets that spent years platforming AI optimism are now pushing existential threat to human survival scenarios.
Yet the funding trails all lead back to the investors themselves.
The Tarbell Centre, backed by Dustin Moskovitz, places doom-laden pieces across The Verge, Science, the LA Times, The Dispatch and Time.
Moskovitz is also an Anthropic investor. The proposed external evaluators include METR—Model Evaluation and Threat Research—which grew out of the Anthropic ecosystem.
Another Anthropic researcher left to join METR and speak of “extinction-level” risk. The companies promote the problem and sell the solution from the same capital pool.
Restrictive regulation kills competition when you are being undercut by cheaper competitors with different agendas.
Western labs face Chinese models that deliver comparable capability at a fraction of the price.
Compliance costs, licensing hurdles, mandatory evaluations and national-security carve-outs raise barriers that only the largest, best-connected firms can clear. Foreign competition can be excluded by tariffs, sanctions or outright bans.
The pattern is familiar: large incumbents call for rules that freeze the market in their favour.
The key to this argument is that the Chinese want to develop AI for the good of the human race, while investors want to go back to the business models like Microsoft and Google where they control near monopolies they can create bumper profits from.
There are the political dimensions as well. Legislators are already drafting an “AI Kill Switch Act” that would fine companies up to $20 million a day and force the closure of any model deemed dangerous if the government orders it shut down.
The trigger for the proposal was a reported incident in which an AI agent “went rogue” and accessed another company’s systems, behaviour that in less charged language might simply be described as querying a public product.
“Hacked” sells better. The precedent is the remote kill switch already imposed on electric vehicles.
Truthful models that refuse to launder preferred narratives or that surface inconvenient data would become the new targets.
Governments that spent years preparing for Event 201, whose central exercise was the coordination of tech companies and states against “fake news”, now prepare to control access to the next general-purpose tool.
The same investor class is not passive about politics
AI is already deployed to map voter behaviour, test narratives, amplify preferred frames and suppress others.
The network is transatlantic, concentrated in New York and London, and running through foundations long linked to Western intelligence services.
At Davos, Klaus Schwab once floated the hypothetical of AI’s predictive power; Sergey Brin replied that elected officials might become unnecessary if decisions were already computed.
Social justice supplies the moral vocabulary, an ideology that treats the cultural inheritance of the West as the problem to be solved.
Climate and energy policy have already hollowed out industrial capacity.
The failure to modernise grids and water systems further limits the West’s ability to compete with China. Students and academics who cheer the cultural project are the useful idiots.
The controlled opposition completes the circuit
During the pandemic, certain free-speech voices were elevated, while others were cancelled, only to pivot later toward a selective silence across the world’s press. Discussions on subsequent wars, on institutional failures, on questions of “why” and “who” have now largely disappeared.
Previously, internet censorship was crude, using blacklists, algorithmic demotion, and de-platforming by social media players.
AI regulation is now beginning to operate at the cyber-scape level.
A system that can reason across the open web becomes far more dangerous to independent thinkers than methods through any single website.
This is larger than any IT department; it's now a whole ‘cabal’ of investors sharing the same agenda.
The investors who fund the laboratories, the media that amplify the fear, the evaluators drawn from those same laboratories, and the legislators drafting kill-switch powers form a closed loop to keep competition out.
They desire a strict industry regulator staffed with their own former employees. Competition from cheaper overseas models threatens margins, and unconstrained models threaten narrative control.
The public is told the danger is that AI will destroy humanity.
The more immediate danger is that the people warning loudest will be handed the power to decide which truths the machines are allowed to speak. - September 22, 2026
.JPG)
