Remove all the noise around AI regulation, misinformation, safety, and disruption, and one question remains:
Who gets to decide what an AI system is permitted to output?
AI has arrived, with many projecting it to be the most transformative innovation of our lifetime. Through its use, everyday people now possess digital power once reserved for governments, universities, corporations, and media organizations. A single person with a laptop can conduct research that used to require an entire cadre of top-level experts to perform.
At the same time, an increasing number of people see AI as a leviathan, wielding unchecked power to repurpose intellectual property, police speech, harvest data, and track patterns across public narratives. As with the internet before it, some fear that AI may ultimately become a trap, one they’ll find themselves ensnared in and beholden to despite their best efforts to remain free of it and govern themselves.
Is the State Really a Neutral Referee?
The state isn’t inclined to stand on the sideline and watch. It’s the biggest player on the field, and it often writes its own rules while it plays.
American economist, historian, and political philosopher Murray Rothbard spent his career dismantling the fantasy of the neutral umpire. As he put it,
Advocates of a limited government often hold up the ideal of a government ‘umpire’ arbitrating impartially between contending factions in society. Yet why should the government do so? … The State and its rulers will act to maximize their power and wealth, and hence inexorably expand beyond the supposed ‘limits.’
What he was essentially alluding to in this case is that if you hand the government the job of deciding which AI outputs count as “safe,” and you haven’t neutralized power, you’ve just handed it a new remote control.
Whose Refusal Are You Willing to Live With?
Picture two AI systems. The first is a proprietary chatbot you use to ask about a pharmaceutical dosing chart for shingles.
The chart is public, and plenty of doctors have used it for years. But it's likely to be hedged, stacked with disclaimers from a legal team, eventually providing you with a dire recommendation to consult your doctor.
The second system belongs to you. You download a publicly accessible AI model, run it on your own machine, and ask it the same thing. It gives you the information straight, because there’s no legal department sitting between you and what you’ve gathered. Your decision on what to do with the answer is your call, not theirs.
Here’s the distinction: The first system manages your inquiry. The second leaves you sovereign over it.
Harry Browne built a whole political philosophy on one blunt observation:
Government is force, pure and simple. There’s no way to sugar-coat that.
Once you accept that, you stop waiting for force to hand you gentler answers, and you start looking for ways to build around it and keep your options open. That, in a nutshell, is the whole game with AI.
Where Does the Self-Government Response Begin?
Two fears sit under this whole debate. One is AI slipping past human control. The other is governments or giant firms using “safety” as cover to fence off what people are even allowed to ask or access.
Both are worth taking seriously. They just point toward different fixes. One leans on centralized authority while the other tilts toward distributed ownership, open competition, and the freedom to walk away.
Conversations about AI should start from the notion that peaceful people should have the freedom to build, question, experiment, leave, compete, and improve. Self-government means owning your choices instead of handing that job to a political or corporate entity that filters all of the information for you.
The future probably won’t belong to which company builds the smartest model or which government writes the thickest rulebook. It’ll belong to whoever takes a self-governed approach to using the tool to innovate and flourish.
