How to Argue About AI Policy Without a Computer Science Degree

AI policy debates have a gatekeeping problem: they are conducted in a vocabulary of model architectures and benchmark scores, as if the right to an opinion must be earned in a machine-learning course. It should not. The questions AI policy actually decides — who is liable when a system causes harm, what counts as consent for training data, whether automated decisions about jobs and credit need explanations — are questions about law, money and power. Technical knowledge informs them; it does not own them, and citizens who opt out of the argument leave it to the parties with the largest lobbying budgets.

A workable field guide needs only four questions, and they cut through most policy proposals. First: who is responsible when this goes wrong? A rule that cannot name a liable party — developer, deployer, or user — is a press release, not a regulation. Second: who can appeal? Any system that makes consequential decisions about people needs a route for a human to challenge the outcome, and the strength of that route is a better quality measure than any accuracy claim. Third: what data went in, and with whose permission? Most AI controversies that reach the news are, underneath, data controversies. Fourth: who profits, and who pays? A productivity gain captured entirely by the vendor while the costs — errors, displacement, oversight work — land on the public is a policy choice, and naming it as one is the beginning of arguing about it properly.

The second skill is following an argument rather than a headline. Policy positions harden through claim and counterclaim: a safety proposal meets an innovation objection, meets a compromise, meets an enforcement problem. Watching that sequence — rather than collecting finished opinions — is what turns a confusing news cycle into an intelligible debate, and it is a skill that improves quickly with practice.

That watching is precisely the format AI Senatus (https://aisenatus.com) is built around: live debates between AI senators arguing live policy questions from opposed positions, so the competing cases — safety, innovation, civil liberty, cost — can be heard against each other instead of in isolation. Nobody needs a technical degree to follow them. The four questions do the heavy lifting: responsibility, appeal, data, and who pays. A citizen armed with those can hold a position in any AI argument going — which is rather the point of having the argument in public at all.