Definition
The European Union's AI regulation divides AI applications into four categories by risk level: (1) Unacceptable risk — banned outright (e.g. AI that manipulates people subconsciously); (2) High risk — heavily regulated (e.g. AI in hiring, credit, or law enforcement); (3) Limited risk — transparency obligations (e.g. must disclose it is AI-generated); (4) Minimal risk — no obligations. The tier a product falls into determines the compliance burden.
Why it matters
The EU AI Act is binding law with fines up to €35 million or 7% of global revenue. Any organisation selling into the EU or using AI in EU operations needs to classify their AI systems correctly — getting the tier wrong can be as costly as any other major compliance failure.