Section 1 of 4
Part 1 named the system’s loops. This section maps your structural position inside them — not as a moral indictment, but as a starting point for knowing which levers you’re closest to.
Everyone who participates in the AI economy occupies a structural position in it. That position determines which leverage points you’re naturally adjacent to, which ones require access you’d need to build, and which are genuinely out of reach without coordination with others. The goal isn’t to produce guilt about where you sit. It’s to produce accuracy — because different positions give access to different levers, and the levers aren’t interchangeable.
There are three positions most people occupy simultaneously.
As a consumer or enterprise user, your subscription or contract sends a revenue signal into R1 — the capability-capital-compute loop. Lab revenue supports compute procurement, which supports training runs, which produces the capability that justifies the price. This is a small position in the overall loop structure: consumer revenue is a fraction of what institutional investors, government contracts, and enterprise deployments generate. But the subscription relationship also creates standing you may not have fully used. As a paying customer, you have standing to demand disclosure in vendor contracts, flag harm to product teams, and apply pressure through procurement — particularly in enterprise contexts where a single contract is materially significant.
As a professional whose work is being redesigned by AI, you’re one of the flows the capability stock is being built to partially substitute. That’s not a call to resist AI at work — it’s a structural observation with a practical implication. Workers who understand which tasks are being substituted versus augmented are better positioned to shape how AI is deployed inside their organizations. The EU AI Act (Article 26) requires employer consultation before deploying high-risk AI in employment contexts; that leverage exists in labor processes even where it’s not being used. In the US, no equivalent federal requirement exists, which means the leverage must be built from contract language, internal policy, and organized professional pressure rather than statute.
As a voter and civic participant, you’re one of the inputs to the slow balancing loop (B) that determines whether oversight catches up to capability. Political conditions aren’t natural facts — they’re feedback signals from elections, public discourse, and where organized attention is directed. The EU AI Act exists because that attention was applied. The US federal equivalent doesn’t because it hasn’t been, not because the problem is structurally different. Your position in this loop is real, operates on a longer timeline than R1, and is the entry point for the highest-leverage structural changes — the ones that require legislative or judicial action.
Your professional position shifts your leverage tier
A software engineer at an AI lab can influence buffer sizes — the slack that determines whether pre-deployment review periods shrink under shipping pressure. A lawyer handling AI liability cases is directly working strength of the brakes. A labor organizer in a sector being automated is working on what the system is actually for. These aren’t career recommendations — they’re structural observations about where professional standing gives you access most people don’t have. The section on reading your leverage maps this more specifically.
None of these positions require that you view AI with alarm or enthusiasm. They’re structural facts about where you sit, which is the starting point for choosing which levers to engage.
Section 2 maps what different kinds of action actually move — and why most public AI safety activity operates at the weakest tier.