EU AI Act for Manufacturing: Product-Safety AI, Machinery & When It Turns High-Risk
Manufacturers deploy AI for predictive maintenance, quality inspection, process optimisation and robotics. Under the EU AI Act (Regulation (EU) 2024/1689), much of this is minimal-risk, but there is a critical exception: AI acting as a safety component of machinery or other regulated products is high-risk through the product-safety route. That route has a later deadline than the Annex III uses, giving manufacturers a longer — but firm — runway.
Is it in scope?
The high-risk trigger here is Art. 6(1): an AI system is high-risk if it is a safety component of a product covered by the EU harmonisation legislation in Annex I (such as the Machinery Regulation) and that product must undergo third-party conformity assessment. AI embedded in the safety function of industrial machinery, robots or equipment can therefore be high-risk, with obligations applying from 2 August 2027. Purely operational uses — predictive maintenance, yield optimisation, demand forecasting — are usually minimal-risk. Where AI is used to manage or monitor the factory workforce, the Annex III(4) employment rules apply instead.
Typical AI use cases
- AI safety components in machinery and robotics
- Predictive maintenance of equipment
- Computer-vision quality inspection
- Production-line and process optimisation
- Worker allocation and productivity monitoring
Risk classification
Two distinct regimes can apply. AI that is a safety component of Annex I machinery or equipment is high-risk under Art. 6(1), with the manufacturer as provider carrying Art. 9–15 duties — integrated where possible into the existing product conformity assessment, effective 2 August 2027. AI used for worker management (task allocation, monitoring) is high-risk under the Annex III(4) employment listing, from 2 August 2026. Everything else — maintenance, quality analytics, optimisation — is generally minimal-risk with no mandatory obligations beyond AI literacy under Art. 4.
Obligations to prepare for
Your exact duties also depend on whether you build or use the AI. See obligations by operator role — provider, deployer, importer, distributor or GPAI provider.
FAQ
Is predictive maintenance AI high-risk?
Generally no. Predictive maintenance, yield optimisation and demand forecasting are not listed in Annex III and are typically minimal-risk. The high-risk trigger arises where AI performs a safety function in a regulated product such as machinery.
When does the product-safety deadline apply?
AI that is a safety component of a product covered by Annex I harmonisation legislation requiring third-party conformity assessment is high-risk under Art. 6(1), with obligations applying from 2 August 2027 — later than the 2 August 2026 date for most Annex III uses.
We use AI to allocate tasks and monitor workers — is that covered?
Yes, but under a different heading. Worker management and monitoring fall under the Annex III(4) employment listing and are high-risk from 2 August 2026, with the employer carrying Art. 26 deployer duties including worker information.
Run the free 2-minute risk checker — no sign-up.