On: Anthropic suspends new AI tools over US government security concerns

The news that Anthropic has suspended its new tools because the US government is concerned - this is a small symptom that reveals the whole architecture. Notice: no law was passed, no democratic debate occurred, no public had its say. A private corporation, in consultation with a state that never needed to issue a formal command, has decided that certain possibilities are simply not thinkable. That is hegemony working as it is meant to: without a visible hand, without a signature, without a protest to be made.

What is the “concern”? Cybersecurity, hacking, the vague threat of misuse. But let us ask cui bono of the concern itself. Who benefits from the assumption that this technology must be surveilled before it can be built? Certainly not the user, who never consented to this horizon being closed. Certainly not the engineer, who now knows that the thinkable is smaller than the feasible. The beneficiary is the state’s monopoly on the imagination of risk - the power to decide which futures we are allowed to test.

This is the interregnum speaking. The old order of state sovereignty over technology is dying - it cannot enforce its will through law alone, because the technology moves too fast. The new order is not yet born - no democratic framework for governing these tools exists. In the gap, we get symptoms: voluntary suspensions, anxious press releases, a corporation that acts as a ministry of the interior without admitting it. The company’s executives do not know they are organic intellectuals of the security apparatus. They believe they are acting responsibly. That is precisely what makes them so effective.

The war of position here is being lost before it has begun - not because anyone has conquered the field, but because the field itself has been declared unplayable.