What is EU AI Act Annex III?

Annex III is the schedule to the EU AI Act that lists the areas where AI systems are high-risk. It covers 8 areas, biometrics, safety components of critical infrastructure, education, employment, essential services (credit, life and health insurance, public benefits, emergency calls), law enforcement, migration and border control, and justice and democratic processes.

Who does Annex III apply to, providers or deployers?

Both, but with different obligations. Providers (companies that build or develop the AI system) carry the heaviest obligations, conformity assessment, technical documentation, CE marking, registration in the EU database.

When must high-risk AI systems comply?

The Digital Omnibus on AI, Regulation (EU) 2026/1744, in force since July 27, 2026, moved the Annex III date by 16 months. Standalone Annex III AI systems must comply by December 2, 2027. High-risk AI in products covered by the EU product laws in Annex I, such as medical devices, toys and lifts, has until August 2, 2028.

What is a conformity assessment for high-risk AI?

A conformity assessment is the process of verifying that your AI system meets all the requirements in Articles 9-15 of the EU AI Act. For most Annex III systems, this is a self-assessment process, the provider conducts the assessment internally, documents the results in technical documentation, and issues an EU Declaration of Conformity.

EU AI Act Annex III High-Risk AI Systems: Checklist

Q: What does the EU AI Act database require?

Providers of high-risk AI systems listed in Annex III must register themselves and the system in the EU database before placing it on the market (Article 49(1)). Critical infrastructure systems are registered at national level instead. The entry covers the provider, the system, its intended purpose and its conformity status (Annex VIII).

Most EU AI Act coverage focuses on the GPAI rules (for model builders) or the Article 50 transparency requirements (for chatbots and deepfakes). The Annex III rules for high-risk AI systems get less attention, but they cover more companies. If your company uses AI to screen job applicants, approve credit, price life or health insurance, or monitor employee performance, you are likely deploying a high-risk AI system under EU law.

The Digital Omnibus on AI, Regulation (EU) 2026/1744, in force since July 27, 2026, moved the Annex III compliance date by 16 months, from August 2, 2026 to December 2, 2027. That is still more than a year away. Enough time to build a compliant program if you start now, not enough time to ignore it.

TL;DR: Annex III of the EU AI Act lists 8 areas of high-risk AI. The Digital Omnibus (Regulation (EU) 2026/1744) pushed the compliance date for standalone Annex III systems to December 2, 2027. Providers must complete conformity assessment, build technical documentation, and register before that date. Deployers must implement human oversight, keep logs, and inform affected workers. Prohibited AI practices (Article 5) have applied since February 2, 2025. Those did not move.

The 8 Annex III categories

Annex III defines high-risk AI by use case, not by technical architecture. The same underlying model can be high-risk in one context and not in another.

EU AI Act Annex III compliance framework overview for high-risk AI systems

1. Biometrics

Annex III point 1 covers three kinds of system, in so far as Union or national law permits their use:

Remote biometric identification: identifying who someone is from their face, gait or other biometric data, at a distance. One-to-one biometric verification (checking that you are who you claim to be, like a passport scan matched to a selfie) is excluded.
Biometric categorisation by sensitive or protected attributes inferred from biometric data. Categorising people to infer race, political opinions, trade union membership, religious or philosophical beliefs, sex life or sexual orientation is banned outright under Article 5(1)(g).
Emotion recognition. Inferring emotions at the workplace or in education is banned under Article 5(1)(f), except for medical or safety reasons.

Real-time remote biometric identification in publicly accessible spaces for law enforcement is banned under Article 5(1)(h), with narrow exceptions. Identifying a person from recorded footage after the fact (post-remote identification) is high-risk, and law enforcement needs an authorisation for it (Article 26(10)).

2. Critical infrastructure (safety components)

AI systems used as safety components in the management and operation of critical infrastructure, including energy grids, water supply, transport networks, and digital infrastructure. Applies when AI failure could endanger health, safety, or essential services at scale.

Most small teams are not in this category. Exceptions: IoT platforms that feed into utility management systems, industrial control software that interfaces with grid operations, or fleet management AI used by transport operators.

3. Education and vocational training

AI systems used to:

Determine access to educational and vocational training institutions
Assess students during examinations (AI proctoring that can determine exam outcomes)
Evaluate learning outcomes in ways that affect access to education
Monitor and detect student behavior (cheating detection systems that trigger consequences)

Not covered: AI tutoring tools, recommendation engines suggesting courses, grammar checkers used for feedback.

4. Employment and worker management

This is the category most relevant to companies outside healthcare and finance. It covers AI used to:

Recruit or select job applicants (resume screening, interview analysis, automated ranking of candidates)
Make or significantly influence decisions about promotion, assignment, or termination
Monitor and evaluate employee performance and behavior
Allocate tasks to workers and determine their conditions of work

The EU AI Act employment category overlaps significantly with US laws: Illinois AIVRA (for video interviews), Colorado SB 26-189 (automated tools in employment and other consequential decisions, from January 1, 2027), and EEOC guidance on AI in hiring. Companies operating in both jurisdictions face compliance obligations from multiple directions.

Most affected tools: HireVue, Pymetrics, Workday AI, SAP SuccessFactors AI modules, Greenhouse AI scoring, any custom-built resume ranking system.

5. Essential private and public services: credit, life and health insurance, benefits

Annex III point 5 lists four uses:

Evaluating creditworthiness or setting credit scores for individuals, with the exception of AI used to detect financial fraud
Risk assessment and pricing for individuals in life and health insurance (car, home and other insurance lines are not in Annex III)
Evaluating eligibility for essential public assistance benefits and services, including healthcare, when used by or on behalf of public authorities
Evaluating and classifying emergency calls, dispatching emergency services, and emergency healthcare patient triage

This covers fintech credit decisioning AI, AI that prices life or health cover, and AI used by banks to approve or price credit for individuals.

6. Law enforcement

AI systems used by law enforcement authorities to:

Assess the risk of an individual offending or reoffending
Conduct risk assessments for use in criminal investigations
Profile individuals in the course of detection, investigation, or prosecution

Applies primarily to public authorities. Private companies building tools for law enforcement agencies are providers of high-risk AI. Pure commercial tools (fraud detection for private banks, private security AI) are generally not in this category unless they feed into law enforcement processes.

7. Migration, asylum, and border control

AI systems used by or on behalf of competent public authorities to:

Act as polygraphs or similar tools
Assess a security, irregular migration or health risk posed by a person entering or already in a Member State
Assist in examining applications for asylum, visas and residence permits, and related complaints
Detect, recognise or identify people (verification of travel documents is excluded)

Applies primarily to public authorities. Vendors building AI for border agencies or immigration authorities are subject to the high-risk rules as providers.

8. Administration of justice and democratic processes

AI used by or for a judicial authority to research and interpret facts and the law and apply the law to facts, or used the same way in alternative dispute resolution. Also covers AI intended to influence the outcome of an election or referendum or how people vote. Tools whose output voters never see directly, such as software that organises a campaign's logistics, are excluded.

Decision table: is your AI high-risk?

If your AI does this...	Annex III category	High-risk?
Screens job applications	Employment (4)	Yes
Monitors employee productivity	Employment (4)	Yes
Scores creditworthiness	Essential services (5)	Yes
Prices life or health insurance for individuals	Essential services (5)	Yes
Prices car or home insurance	Not listed	No
Identifies faces in real time in public spaces for the police	Biometric (1), Article 5(1)(h)	Banned, narrow exceptions
Matches faces to a database after the event	Biometric (1)	Yes
Determines access to a university course	Education (3)	Yes
Proctors online exams	Education (3)	Yes
Recommends courses to learners	Education (3)	No
Detects fraud for a private bank	Not listed	No
Evaluates the reliability of evidence for the police	Law enforcement (6)	Yes
Drafts emails using AI	None	No
Summarizes legal documents internally	None	No
Assists a judge in researching case law	Justice (8)	Yes
Optimizes energy usage in a building	None	No
Is a safety component in running a power grid	Critical infrastructure (2)	Yes

What providers must do before December 2, 2027

Providers are companies that develop and place high-risk AI systems on the EU market. This includes software vendors who sell AI tools used in high-risk contexts, even if they do not do the final deployment themselves.

1. Risk management system (Article 9) Implement and maintain an ongoing risk management process for the AI system. This is not a one-time risk assessment but a continuous process that must run through the system's lifetime.

2. Data governance (Article 10) Training, validation, and testing data must meet quality criteria. Data must be relevant, representative, and free of errors that could introduce bias affecting protected groups. You must document the data sources and the measures taken to identify and address biases.

3. Technical documentation (Article 11 + Annex IV) Prepare documentation covering: general description of the system, intended purpose, elements of the system, changes made, system performance metrics, datasets used, risk management, and human oversight measures. This documentation must be ready for regulators on request.

4. Automatic logging (Article 12) High-risk AI systems must generate logs automatically. The logs must enable post-hoc monitoring of the system's operation. Providers and deployers each keep the logs under their control for at least six months (Articles 19 and 26(6)).

5. Transparency and instructions for use (Article 13) Provide deployers with clear instructions covering: the identity of the provider, the capabilities and limitations of the system, the performance metrics and their test conditions, any known risks, the human oversight measures needed, and the technical measures to support oversight.

6. Human oversight (Article 14) Design the system so humans can effectively oversee it. This means the system must allow human intervention, must not prevent humans from overriding outputs, and must communicate to users when the system cannot reliably perform. Oversight is not optional. If your system is designed to bypass or minimize human review, it is non-compliant.

7. Accuracy, robustness, cybersecurity (Article 15) The system must achieve appropriate accuracy for its intended purpose, be robust against errors and attempts to manipulate outputs, and meet cybersecurity requirements proportionate to the risks.

8. Conformity assessment Annex III points 2 to 8 use the internal-control procedure, with no notified body (Article 43(2)). For point 1 (biometrics), a notified body is involved where the provider has not fully applied harmonised standards or common specifications (Article 43(1)). Issue an EU Declaration of Conformity. Affix CE marking.

9. Register in the EU database Register yourself and the system in the EU database before placing it on the EU market (Article 49(1)). Critical infrastructure systems (point 2) are registered at national level instead (Article 49(5)).

What deployers must do before December 2, 2027

Deployers are companies that use a high-risk AI system in a professional context in the EU, even if they did not build it.

1. Fundamental rights impact assessment (FRIA) Deployers that are public bodies or private entities providing public services, and any deployer using AI for credit scoring or for risk assessment and pricing in life and health insurance, must conduct a fundamental rights impact assessment before deployment (Article 27). Private employers using hiring AI are not required to conduct a formal FRIA, but documenting the risk review is strongly recommended.

2. Human oversight Assign qualified persons to oversee the AI system. Ensure they have the authority and practical ability to override or stop outputs. Train them in the system's capabilities and known limitations.

3. Maintain use logs Deployers must retain the logs generated by the AI system for at least 6 months (longer if the sector requires it; financial records and HR records often impose longer retention). These logs are subject to regulatory inspection.

4. Inform workers Before you put a high-risk AI system into use at the workplace, inform workers' representatives and the affected workers (Article 26(7)).

5. Data protection For high-risk AI systems that process personal data, your existing GDPR obligations apply. The AI Act and GDPR obligations run in parallel. Article 22 GDPR rights to human review of automated decisions overlap with the AI Act's human oversight requirement. Maintain both.

6. Register (if applicable) Deployers that are public authorities must register their use of an Annex III system in the EU database, except for critical infrastructure systems (Article 49(3)).

Transitional period: what the EU Digital Omnibus changed

The Digital Omnibus on AI, Regulation (EU) 2026/1744, published on July 24, 2026 and in force since July 27, changed the timeline:

Standalone Annex III AI systems (employment, credit, education, biometrics and the other Annex III areas): moved from August 2, 2026 to December 2, 2027.
High-risk AI in Annex I products (medical devices, toys, lifts and the other product laws in Annex I): moved from August 2, 2027 to August 2, 2028.
Prohibited AI practices (Article 5): apply since February 2, 2025, unchanged. The Omnibus added two bans (non-consensual intimate imagery and child sexual abuse material) that apply from December 2, 2026.
GPAI obligations (Chapter V): applied since August 2, 2025, unchanged.

Systems already on the market get a transitional rule (Article 111(2)). A high-risk system placed on the market or put into service before the Annex III date is covered only if its design changes significantly from that date. Systems intended for public authorities must comply by August 2, 2030 in any case.

If your system is already deployed and you plan no significant design changes, document that position now so you have a defensible record when national authorities start inspecting.

Compliance checklist for December 2, 2027

For providers:

For deployers:

Inventory all AI systems in use and identify which fall under Annex III
Request technical documentation and instructions for use from each provider
Assign qualified persons with authority to oversee each high-risk system
Implement the human oversight measures specified by the provider
Configure log retention for at least 6 months
Inform workers' representatives and affected workers before using a high-risk system at the workplace
Complete a fundamental rights impact assessment if required (public bodies, private entities providing public services, credit scoring, life and health insurance pricing)
Verify your GDPR Article 22 compliance runs in parallel with AI Act oversight obligations
Register if you are a public body deploying Annex III AI

Enforcement and penalties

The EU AI Act assigns enforcement to national market surveillance authorities in each member state. The EU AI Office coordinates at the European level, particularly for GPAI and systemic risk cases.

For violations of the Annex III obligations, fines reach up to 15 million EUR or 3% of global annual turnover, whichever is higher (for SMEs and start-ups, whichever is lower). Deploying a prohibited AI system (Article 5) carries higher penalties: up to 35 million EUR or 7% of global turnover.

Enforcement timelines by obligation type:

Prohibited practices (Article 5): enforceable since February 2, 2025
GPAI obligations: August 2, 2025 (Commission fines on GPAI providers from August 2, 2026)
Annex III standalone high-risk systems: December 2, 2027
High-risk AI in Annex I regulated products: August 2, 2028

The current enforcement phase (February 2025 onward) focuses on the Article 5 prohibitions: social scoring AI, real-time remote biometric identification by police in public spaces, AI that exploits vulnerabilities, and AI that manipulates behavior subliminally. If your system touches any of those, it is already in scope. No deadline extension applies.

Correction (October 16, 2026): Earlier versions cited Colorado SB 205's high-risk AI rules. That law never took effect; SB 26-189 replaced it for decisions from January 1, 2027 (signed act).

Correction (October 16, 2026): Earlier versions dated the Digital Omnibus to a May 7, 2026 provisional agreement we could not verify and said GPAI obligations start August 2, 2026. The delay is law under Regulation (EU) 2026/1744, and GPAI obligations have applied since August 2, 2025 (Regulation (EU) 2024/1689).

Correction (October 17, 2026): Earlier versions listed all insurance risk assessment as Annex III high-risk and limited the fundamental rights impact assessment for credit and insurance to deployers acting for public bodies. Annex III point 5(c) covers only life and health insurance, and Article 27 applies to every deployer of point 5(b) and 5(c) systems (consolidated AI Act).

Correction (October 19, 2026): A check of every statement against the consolidated AI Act found errors the October 16 and 17 corrections missed. The body still described the Omnibus as the May 7 provisional agreement, not yet law, and gave the old Annex III date as December 2, 2027; it is law since July 27, 2026 and moved the date from August 2, 2026. The May 7, 2026 provisional agreement itself did happen (European Parliament research service); the October 16 note should not have called it unverified. We had listed insurance claims triage and insurance premiums in general (only life and health insurance pricing is in Annex III), travel document checks (excluded from point 7), access control by face as identification (one-to-one verification is excluded), a third-party assessment for critical infrastructure AI (only point 1 biometrics can need a notified body), registration for public deployers in "categories 1-6 and 8" (every area except point 2), and GPAI obligations from August 2, 2026 (they applied from August 2, 2025).