The AI regulations cover four categories of risk; unacceptable risk, high risk, limited risk and minimal/ no risk.