Responsible AI Platform
Safety Components & CE

AI Act Compliance for Manufacturing & Industry

Robotics, quality control and safety components — regulated under the AI Act

Practical guidelines for manufacturing companies and industrial organizations to comply with the EU AI Act and Machinery Regulation.

View the compliance checklist

Why Take Action Now?

The AI Act has major impact on the manufacturing sector

August 2025

First obligations for AI as safety component in machines and production systems

Safety Component = High-risk

AI affecting machine safety falls under the strictest AI Act and Machinery Regulation

Fines up to €35 million

Or 7% of global annual turnover — plus CE marking withdrawal for non-compliance

CE Marking Obligation

AI safety components require conformity assessment and European market access

High-risk AI in Manufacturing & Industry

These AI applications fall under strict AI Act requirements

Robotics & Cobots

Industrial robots and collaborative robots with AI control — worker safety is paramount.

Autonomous robotsCobot control systemsRobot path planningCollision avoidance AI

Quality Control AI

Vision AI for product inspection and quality monitoring — impact on product liability and safety.

Visual inspection AIDefect detectionDimensional controlSurface quality analysis

Predictive Maintenance

AI predicting machine maintenance and preventing production downtime — safety-relevant for critical components.

Machine health monitoringRemaining useful lifeVibration analysis AIThermal monitoring

Process Optimization & Safety

AI systems optimizing production processes with direct impact on worker and product safety.

Process parameter optimizationEnergy consumption AISupply chain optimizationEmissions monitoring

Specific Challenges for Manufacturing

The AI Act brings unique compliance questions for the industrial sector

Machinery Regulation Integration

How does AI Act compliance fit into existing CE marking processes? What are the overlaps with the new Machinery Regulation?

Safety Component Definition

When is AI a "safety component"? The classification determines whether you must meet the strictest requirements.

OT/IT Convergence

Operational technology with AI requires a different approach than office IT. How do you document embedded AI in machines?

Legacy Machines with AI Retrofit

Existing machines are increasingly equipped with AI. Who is responsible for compliance with retrofits?

Supply Chain Responsibility

Complex supply chains with multiple AI components. Who is provider, who is deployer?

Workplace Safety & AI

AI monitoring or protecting workers — where is the line between safety and surveillance?

AI Act Compliance Roadmap

Practical steps for manufacturing companies

1

AI Inventory

2-4 weeks

Map all AI systems in machines and production processes. Including embedded AI and retrofits.

2

Risk Classification

1-2 weeks

Determine per system whether it is a safety component and whether it falls under high-risk.

3

Gap Analysis

3-6 weeks

Compare current CE documentation and safety processes with AI Act and Machinery Regulation.

4

Remediation

3-12 months

Implement technical documentation, risk management, testing and conformity assessment.

5

Ongoing Monitoring

Ongoing

Set up processes for post-market surveillance and periodic safety reviews.

18-month trajectory

Implementation Roadmap

Detailed 6-phase trajectory with concrete deliverables

1

Inventory

Month 1-2
Complete AI system register (production + logistics)Owners per systemLink with existing machine register
2

Classification

Month 2-3
High-risk vs. limited/minimal per systemSafety component assessment (Annex I)Annex III mapping for production AI
3

Gap Analysis

Month 3-5
Gap per system: AI Act + Machinery RegulationCE documentation reviewPrioritization based on risk and deadline
4

Governance Framework

Month 5-7
AI governance structureRoles & responsibilitiesIntegration into ISO 9001 quality management
5

Implementation

Month 7-14
Technical documentation per systemHuman oversight mechanismsConformity assessment trajectories startSupplier assessment
6

Audit-ready

Month 14-18
Internal auditDry-run for notified bodyPost-market surveillance operational

AI System Inventory

Typical AI systems in manufacturing and their likely classification

Note: many PLCs and embedded controllers now contain AI functionality without it being explicitly labeled as such. Also inventory AI in retrofits and supplier software.

Robotics & Cobots

Usually high-risk
Autonomous navigationCollision avoidancePick & place AICobot motion planning

Annex I + III — AI control of robots is a safety component under the Machinery Regulation

Quality Control & Inspection

Context-dependent
Vision AI defect detectionDimensional controlSurface qualityX-ray/CT inspection AI

High-risk if the product is a safety component; otherwise often limited risk

Predictive Maintenance

Context-dependent
Vibration analysisThermal monitoringWear predictionOil analysis AI

High-risk if failure poses safety risk; limited risk for non-critical machines

Autonomous Vehicles & AGVs

Usually high-risk
Warehouse AGVsAutonomous forkliftsYard managementIndoor navigation

Annex III — autonomous vehicles in work environments are high-risk due to worker safety

Worker Safety

Often high-risk
PPE detectionHazard zone monitoringErgonomics analysisExposure monitoring

High-risk when AI makes autonomous safety decisions about workers

Supply Chain & Planning

Usually minimal risk
Demand forecastingInventory managementSupplier assessmentProduction planning

Minimal risk unless it makes autonomous decisions affecting individuals

Classification Decision Tree

Quickly determine the risk classification of your AI system

Is the AI system a safety component of a machine or product (Annex I)?

Yes

Automatically high-risk + CE marking required

No

Go to next question

Does the AI system control autonomous robots, cobots or vehicles in the production environment?

Yes

High-risk — worker safety at stake

No

Go to next question

Does the AI system make autonomous decisions about workers (monitoring, assessment, access)?

Yes

Probably high-risk (Annex III, point 4)

No

Go to next question

Is it supporting with human override and no impact on physical safety?

Yes

Possibly limited or minimal risk

No

Consult an expert for classification

This is a simplified decision tree. Consult your legal team for definitive classification.

Governance Structure

Recommended organizational structure for AI governance in manufacturing

Board of Directors / Management
AI Governance Committee (Production + Engineering + Quality + Compliance)
AI System Owners per production line
Quality & Safety Engineering
Compliance & Regulatory Affairs
Internal Audit & Risk Management

Link AI governance to your existing ISO 9001 quality management system and CE procedures — don't build from scratch.

Key Roles

AI System Owner

Responsible per AI system for compliance, performance and safety documentation

Safety & CE Engineer

Ensures integration of AI Act requirements into CE marking and safety assessments

Human Oversight Officer

Oversight for high-risk systems — adapted for production environments

Supplier Compliance Lead

Assesses AI components from suppliers for AI Act conformity

Compliance Checklist for Manufacturing AI

Concrete checkpoints for each high-risk AI system

AI system registered in EU databaseArt. 49
Risk management system set up (incl. physical safety risks)Art. 9
Data governance & data quality ensuredArt. 10
Technical documentation complete (incl. CE dossier integration)Art. 11
Logging & traceability configured for production environmentArt. 12
Transparency to operators and maintenance personnelArt. 13
Human oversight arranged (emergency stop, override, dashboards)Art. 14
Accuracy, robustness & cybersecurity testedArt. 15
Conformity assessment completed with notified bodyArt. 43
FRIA completed as deployer of high-risk systemArt. 27
Post-market surveillance plan established and operationalArt. 72

This checklist applies per high-risk system. Combine with Machinery Regulation conformity requirements for efficiency.

Common Mistakes

Avoid these pitfalls in AI Act implementation

Overlooking embedded AI

PLCs, controllers and sensors increasingly contain AI functionality. These are often not included in the inventory.

Machinery Regulation and AI Act as separate tracks

Both regulations overlap significantly for safety components. Integrate CE and AI Act compliance for efficiency.

Not inventorying supplier AI

Machine software often contains third-party AI. As deployer you are co-responsible for compliance.

Only involving IT department

Production engineers, quality managers and safety officers must participate from day one.

Forgetting retrofits

AI added to existing machines after the fact also falls under the AI Act. The retrofitter becomes the provider.

Underestimating worker monitoring

AI systems for worker safety monitoring also touch GDPR and labor legislation. Involve the works council early.

What Makes Manufacturing AI Different?

Sector-specific considerations

Machinery Regulation + AI Act

Manufacturing AI falls under two legislative frameworks with their own conformity requirements

Physical Safety

AI decisions have direct physical impact on workers and products

CE Marking Requirement

Without AI Act compliance no CE marking — no access to the European market

Post-market Surveillance

Manufacturers must continue to monitor AI systems for safety after market introduction

Sector-specific regulation

Regulatory Overlap

How the AI Act connects with existing manufacturing regulation

Machinery Regulation (EU) 2023/1230

Overlap: Safety components, CE marking, conformity assessment, technical documentation

Practical tip: One integrated CE dossier covering both Machinery Regulation and AI Act saves double work

General Product Safety Regulation (GPSR)

Overlap: Product safety, market surveillance, recall procedures

Practical tip: AI Act post-market surveillance can connect to existing RAPEX/Safety Gate notifications

REACH & CLP Regulation

Overlap: AI in chemical process monitoring, exposure assessment

Practical tip: AI-driven exposure monitoring can support both REACH and AI Act compliance

GDPR

Overlap: Worker monitoring, biometrics, camera surveillance on the shop floor

Practical tip: FRIA can partially overlap with DPIA — combine where possible and involve the works council

Occupational Health & Safety

Overlap: Worker safety, risk inventory (RI&E), hazardous substances

Practical tip: Integrate AI Act safety risks into existing occupational risk inventory for efficiency

Ready to Start AI Act Compliance?

Practical tools and guidance for manufacturing companies and industrial organizations

Free 30-minute orientation call

or

Practical updates, no legal jargon