GOP rep terrified after ‘closed-door demo’ of new tech 'showed how to kidnap a lawmaker'

The People Most Afraid of AI Are Still Trying to Control Everyone ElseWhat the story actually saysA House Republican says a closed-door demo of Anthropic’s latest model left him rattled: the system could allegedly find a bank vulnerability, empty accounts, then fix the flaw; it could also outline how to kidnap a lawmaker in seconds, or manipulate critical infrastructure like gas pipelines and water treatment. The administration has blocked public release for now, while Trump’s own executive order has also limited states’ ability to regulate AI on their own.Fear Is Not the PointThe headline fact is not that one lawmaker was scared. The fact is that major institutions are treating a technology with obvious abuse potential as something to be managed behind closed doors by the same political class that cannot be trusted to regulate it in public.That is the power structure: private companies demo dangerous capability to lawmakers; lawmakers react with alarm; the public gets told to wait while the same federal apparatus that cannot govern the technology also preempts state attempts to do so. The people with the least direct exposure to the consequences are the ones deciding the rules.Who Enabled the RiskIf the model can be used to attack banks, infrastructure, or a lawmaker, the relevant actors are not abstract “bad actors.” The enabling decisions sit with the company building and demonstrating the model, and with the government structure that concentrates authority while blocking local regulation.Garbarino’s comments do something revealing: they locate danger in the model’s power, but they also expose how little institutional understanding exists among the people nominally responsible for oversight. “95% of my colleagues don’t understand what the hell’s going on” is not a quirky aside. It is an indictment of a political class presiding over systems it cannot competently govern.The Misdirection ProblemThe frame here can easily drift into spectacle: a scary AI, a shocked congressman, some future threat lurking in the machine. That is convenient because it shifts attention away from the actual human choices already made.The model did not appear in a vacuum. Anthropic built it. The administration blocked public release for now. Trump simultaneously narrowed states’ authority to regulate the technology. Those are decisions, not accidents. Treating the story as a mysterious AI problem rather than a record of elite control and regulatory sabotage makes the danger look technical when it is political.This Is a Familiar PatternThe pattern is simple: private power develops a risky system, federal officials centralize the authority to handle it, and state governments are told to stand down. Then the public is asked to accept that the adults are in the room.That is how institutional cowardice works in practice. It hides behind expertise, closed-door briefings, and emergency language while preserving the same basic arrangement: the people most responsible for the hazard keep the power, and everyone else inherits the consequences.The Real StoryThis is not a story about one AI demo terrifying one congressman. It is a story about a governing class that is already outmatched by the tools it has allowed into the world, then reaches for centralization instead of accountability.The deeper political meaning is blunt: when dangerous technology collides with weak oversight, the result is not caution. It is concentrated authority, public dependence, and a shrinking ability to push back before the damage is normalized. Get full access to Systemic Error at paulstsmith.substack.com/subscribe