Risk tiers · Art. 5, 6, 50, 51
Five tiers, one ladder.
A system sits on one rung of the ladder; transparency duties and general-purpose model duties stack on top of it.
01Prohibited practices
Art. 5Practices the Act bans outright. Placing on the market, putting into service or using such a system is prohibited regardless of role.
What falls under it
- Subliminal, manipulative or deceptive techniques that distort behaviour and cause significant harm (Art. 5(1)(a))
- Exploiting vulnerabilities of age, disability or social or economic situation (Art. 5(1)(b))
- Social scoring leading to detrimental or disproportionate treatment (Art. 5(1)(c))
- Predicting criminal offences based solely on profiling or personality traits (Art. 5(1)(d))
- Untargeted scraping of facial images to build recognition databases (Art. 5(1)(e))
- Emotion recognition in the workplace and in education, except for medical or safety reasons (Art. 5(1)(f))
- Biometric categorisation inferring race, political opinions, trade union membership, religion, sex life or sexual orientation (Art. 5(1)(g))
- Real-time remote biometric identification in public spaces for law enforcement, outside the listed exceptions (Art. 5(1)(h))
- From 2 December 2026: non-consensual intimate deep fakes (Art. 5(1)(ba)) and child sexual abuse material (Art. 5(1)(bb))
02High-risk: Annex I products
Art. 6(1), Annex IAn AI system that is a safety component of a product, or is itself a product, covered by the Union harmonisation legislation in Annex I Section A and required to undergo a third-party conformity assessment under that legislation.
What falls under it
- Machinery, toys, lifts, pressure equipment, recreational craft, cableways
- Medical devices and in-vitro diagnostic medical devices
- Radio equipment, gas appliances, personal protective equipment
- Annex I Section B products (civil aviation, vehicles, marine equipment, rail, agricultural vehicles) follow their own sectoral legislation (Art. 2(2))
03High-risk: Annex III areas
Art. 6(2), Annex IIIAn AI system used in one of the eight areas listed in Annex III. Subject to the Art. 6(3) derogation, and always high-risk where the system performs profiling of natural persons.
What falls under it
- Eight areas: biometrics; critical infrastructure; education and vocational training; employment; access to essential services; law enforcement; migration, asylum and border control; administration of justice and democratic processes
- Requirements of Art. 9 to 15, provider obligations of Art. 16 to 22, deployer obligations of Art. 26 and 27
- Applies from 2 December 2027
04Transparency obligations
Art. 50Systems that interact with people, generate synthetic content, perform emotion recognition or biometric categorisation, or produce deep fakes carry disclosure and marking duties, in addition to any other tier.
What falls under it
- Chatbots and voice agents must tell people they are interacting with an AI system (Art. 50(1))
- Generated audio, image, video and text must be marked in a machine-readable format (Art. 50(2))
- Deployers inform persons exposed to emotion recognition or biometric categorisation (Art. 50(3))
- Deployers disclose deep fakes and AI-generated text published on matters of public interest (Art. 50(4))
- Applies from 2 August 2026; existing generative systems by 2 December 2026 (Art. 111(4))
05General-purpose AI models
Art. 51 to 56Obligations for providers of general-purpose AI models, with additional duties for models that present a systemic risk (presumed above the Art. 51(2) compute threshold).
What falls under it
- Technical documentation, downstream information, copyright policy, training-content summary (Art. 53)
- Model evaluation, risk mitigation, incident reporting and cybersecurity for systemic-risk models (Art. 55)
- Applies from 2 August 2025; models placed on the market before that date by 2 August 2027 (Art. 111(3))
06Minimal risk
Art. 4, Art. 95Every other AI system. The Act imposes no system-specific requirements; the organisation still owes the Art. 4 literacy duty and may adopt voluntary codes of conduct.
What falls under it
- Spam filters, demand forecasting, internal coding assistants, inventory optimisation
- Keep the system in the inventory: a change of purpose can change the tier
The eight Annex III areas
A system intended for one of these areas is high-risk unless the Art. 6(3) derogation applies, and always high-risk when it performs profiling.
Point 1
Biometrics
Remote biometric identification; biometric categorisation by sensitive attributes; emotion recognition (where not prohibited by Art. 5)
Point 2
Critical infrastructure
Safety components in the management and operation of critical digital infrastructure, road traffic, and the supply of water, gas, heating or electricity
Point 3
Education and vocational training
Admission or assignment; evaluating learning outcomes; assessing the appropriate level of education; monitoring prohibited behaviour during tests
Point 4
Employment, workers' management and access to self-employment
Recruitment and selection (targeted job ads, filtering applications, evaluating candidates); decisions on promotion, termination or task allocation; monitoring and evaluating workers
Point 5
Access to essential private and public services and benefits
Eligibility for public assistance benefits; creditworthiness and credit scores (except fraud detection); risk assessment and pricing in life and health insurance; emergency call triage and dispatch
Point 6
Law enforcement
Assessing the risk of becoming a victim; polygraphs; evaluating the reliability of evidence; assessing the risk of offending or re-offending (not solely on profiling); profiling in detection and investigation
Point 7
Migration, asylum and border control
Polygraphs; assessing security, irregular-migration or health risks of entrants; examining applications for asylum, visa or residence permits; identifying persons
Point 8
Administration of justice and democratic processes
Assisting judicial authorities in researching and interpreting facts and law; influencing the outcome of an election or referendum or voting behaviour
The Art. 6(3) derogation
An Annex III system is not high-risk where it does not pose a significant risk of harm to the health, safety or fundamental rights of natural persons, including by not materially influencing the outcome of decision making. That is the case where one of the following conditions is met:
One of the following conditions must be met
- (a)the AI system is intended to perform a narrow procedural task
- (b)the AI system is intended to improve the result of a previously completed human activity
- (c)the AI system is intended to detect decision-making patterns or deviations from prior patterns and is not meant to replace or influence the previously completed human assessment without proper human review
- (d)the AI system is intended to perform a preparatory task to an assessment relevant for the purposes of the use cases listed in Annex III
Profiling rule
An AI system referred to in Annex III is always considered high-risk where it performs profiling of natural persons. The profiling rule overrides every condition above.
Documentation and registrationA provider relying on the derogation documents its assessment before placing the system on the market or putting it into service (Art. 6(4)) and registers the system under Art. 49(2).
Annex I: the product route
Annex I lists Union harmonisation legislation. Section A products (for example machinery, toys, lifts, medical devices, radio equipment) enter the high-risk tier when the AI system is a safety component or the product itself and third-party conformity assessment is required. Section B products (aviation, vehicles, marine equipment, rail) stay under their sectoral acts, which integrate the Act's requirements (Art. 2(2)).
Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744; dates per Art. 113 and Art. 111. Corpus snapshot 02024R1689-20260727. Information, not legal advice.
Start with one AI system.
Run the free check, keep the result as your first inventory entry, and invite your team when you are ready.