The manufacturing AI governance landscape
Manufacturing operations are among the most AI-dense environments in the economy — AI is used in design optimisation, production planning, quality control, predictive maintenance, worker safety monitoring, supply chain management, and increasingly in autonomous production systems. This density of AI use, combined with the high-stakes physical environment of manufacturing operations, creates a complex governance landscape that most manufacturing boards have not fully mapped.
The governance complexity has three dimensions. First, AI embedded in products creates product liability and product safety regulation obligations that are distinct from operational AI governance. Second, AI used in manufacturing operations — including AI that affects workers — creates occupational health and safety obligations alongside employment law obligations. Third, AI used in regulated manufacturing sectors (pharmaceutical, aerospace, food safety, nuclear) creates sector-specific safety case obligations that are more demanding than general AI governance requirements.
AI in products: the Annex I intersection
The EU AI Act's Annex I lists the product safety legislation to which AI-embedded products must conform — including the Machinery Regulation, the Radio Equipment Directive, the Medical Devices Regulation, and others. When AI is embedded in a product that falls under one of these Annex I frameworks, the AI system is subject to the requirements of both the product-specific framework and the EU AI Act. For manufacturers, this means the conformity assessment required for the product may need to be extended to specifically address the AI component, and the technical documentation for the product must include AI-specific documentation.
The practical implication for manufacturers: if your product includes an AI component that performs a safety function, the AI system needs to be treated as a safety-critical component subject to the relevant safety case methodology. For automotive manufacturers, this intersects with UNECE regulations on automated driving systems. For machinery manufacturers, it intersects with the Machinery Regulation's essential health and safety requirements. For medical device manufacturers, it intersects with the MDR's clinical evaluation and post-market surveillance requirements.
Worker monitoring AI: the employment governance dimension
Manufacturing facilities increasingly use AI to monitor worker performance, safety compliance, and productivity. These systems — whether camera-based vision AI assessing production line activity, wearable sensor AI monitoring physical load and fatigue, or algorithmic systems managing task allocation and performance metrics — are subject to the employment provisions of the EU AI Act's Annex III. A worker monitoring system used to evaluate performance, make decisions about task allocation, or generate data used in employment decisions is high-risk AI requiring conformity assessment.
