About
Built by practitioners working at the intersection of AI, data, and regulatory compliance — after seeing how many companies genuinely don't know whether their AI product falls under the EU AI Act, a US state law, or both. Not because they don't care, but because the actual answer is scattered across dozens of statutes, agency guidance documents, and fast-moving state bills.
The goal here is narrow on purpose: turn that scattered answer into something checkable in minutes, and something a lawyer can actually review in one sitting — not a replacement for counsel, a running start for the conversation with them.
The classification logic isn't a generic checklist — it's built as a gated decision tree matching the actual structure of the EU AI Act (scope → prohibited practices → high-risk → transparency → general-purpose AI → minimal risk), plus the US state laws that overlap or diverge from it.
Before it's ever run against a real client, every rubric is stress-tested against synthetic worked examples — realistic company scenarios designed specifically to hit the edge cases (a system profiling location data, a company fine-tuning its own model, a case where the law itself has a genuine gap) — and every gap that testing finds gets fixed before it reaches a real assessment.
Click any country to see what applies there. Every EU/EEA country gets the same answer today — the EU AI Act and GDPR don't vary by member state in this rubric — while the United States genuinely differs state by state, which is why it gets its own map below.
⚖ Most of these laws apply based on where your users or output are, not where your company is based. A company in Dubai selling into the EU still needs to comply with the EU AI Act — same for any company, anywhere, serving EU or covered-US-state users.
Click any country to see what's known about its AI/privacy regulation.
Click any state. Six have a specific AI-related law in this rubric today; the rest honestly say so rather than pretending coverage that isn't there yet.
Click any state to see what its law actually cares about.
Every classification cites the specific fact and the specific statute — the same reasoning trail a real assessment has to show its work with.
If an answer is ambiguous, the result treats it as the more cautious reading rather than assuming the best case for you.
Some questions don't map cleanly onto any law's current text — when that happens, it says so explicitly instead of quietly defaulting to a clean pass.
A classification is only as good as the facts behind it — the toolkit is built to say when something needs re-checking, not just when it needs fixing.