US vs EU — who regulates robot safety?

The question "who regulates robot safety" doesn't have one answer. It depends on what kind of robot and which market you're selling into. The regulatory landscape for robots is fragmented across categories that don't talk to each other — and the differences between the US and EU approaches are larger than most people realize.

Industrial robots. This is where the gap is widest. In the EU, the Machinery Regulation 2023/1230 requires pre-market conformity assessment for all machinery, including robots. For high-risk categories listed in Annex I Part A — which now includes machines with AI-based safety functions — a notified body must be involved. CE marking is mandatory before the machine can be placed on the market.

In the US, there is no equivalent pre-market regulation for industrial robots. OSHA sets workplace safety standards and can inspect facilities after deployment, but there is no federal approval required before an industrial robot enters an American factory. The manufacturer is not required to submit to any regulatory body for clearance. Standards like ANSI/RIA R15.06 exist, but compliance is voluntary.

This is not a subtle difference. It means that an industrial robot that would require notified body assessment in Europe can be shipped and installed in the US with no pre-market safety gate at all.

Medical and surgical robots. Here, both markets require pre-market clearance. In the US, the FDA regulates surgical robots as medical devices. A 2024 systematic review published in Nature Digital Medicine found that about 90% of FDA-cleared surgical robots received 510(k) clearance as Class II devices. The remaining were approved through the De Novo pathway. Most remain at Level 1 autonomy — directly controlled by the surgeon.

In the EU, surgical robots fall under the Medical Device Regulation (EU MDR 2017/745). CE marking requires notified body assessment, clinical evaluation, and post-market surveillance. The philosophical approach is similar to the FDA's: pre-market, evidence-based, risk-proportionate.

Autonomous vehicles. In the US, NHTSA oversees vehicle safety through Federal Motor Vehicle Safety Standards (FMVSS). The system is based on manufacturer self-certification — the manufacturer declares compliance, and NHTSA can investigate or recall after the fact. There is no pre-market approval. In the EU, vehicles require type-approval under UNECE regulations, which involves third-party testing before the vehicle can be sold.

The pattern. Across all three categories, the EU tends toward pre-market, third-party assessment. The US tends toward post-market enforcement or manufacturer self-certification. For medical devices, the approaches converge. For industrial robots, they diverge dramatically. For autonomous vehicles, they diverge in mechanism but not necessarily in outcome.

The practical implication for any company building Physical AI systems: if you sell in both markets, your regulatory compliance strategy is not one strategy. It's two. And the gap between them is widest in the category growing fastest — industrial robots with AI-based control.