AI in government: what applies to the public sector?
Adopted 2026-06-20 · ≈ 1 min read · 
edited by Dirk Baaijen
AI that determines access to public services or benefits, or is used in law enforcement, migration or justice, is high-risk under the AI Act. As deployers, public bodies must often carry out a fundamental rights assessment (FRIA) and be transparent to citizens.
Short answer: For government, AI is regulated more strictly. AI that determines who gets access to essential public services or benefits, or that is used in law enforcement, migration/asylum/border control or justice, falls under Annex III and is high-risk. As deployers, public authorities must also carry out a fundamental rights assessment (FRIA, Art. 27) before use, and be transparent to citizens.
Why government is a special case
Government exercises authority over citizens, who often cannot "walk away" to another provider. That is why Annex III designates public uses as high-risk:
- AI that assesses access to essential public services and benefits (such as eligibility for a benefit or allowance);
- AI in law enforcement (risk assessment, evidence evaluation);
- AI in migration, asylum and border control;
- AI supporting the administration of justice.
The fundamental rights assessment is mandatory here
Unlike for many private parties, the FRIA (fundamental rights impact assessment, Art. 27) is an obligation, not an option, for public authorities deploying high-risk AI. Before use you map the impact on fundamental rights and the measures you take.
The lesson of SyRI
The Netherlands has the cautionary example: the SyRI welfare-fraud risk system was struck down by a court in 2020 as breaching Article 8 ECHR for lack of transparency and proportionality (see the enforcement tracker). The childcare-benefits scandal earned the tax authority a GDPR fine. Government AI rarely fails on the technology and often on transparency, proportionality and non-discrimination.
What to do
- Inventory and classify your AI uses against Annex III.
- Do the FRIA before deploying high-risk systems.
- Ensure human oversight, transparency and explainability towards citizens.
- Test for non-discrimination and proportionality — the most common pitfalls.
- Arrange AI literacy for civil servants.
In government, trust counts double: AI that affects citizens falls under the heaviest regime, with a mandatory fundamental rights assessment. Start there, not with the technology.
Sources
- https://eur-lex.europa.eu/eli/reg/2024/1689/oj
Regulation (EU) 2024/1689 (AI Act): Annex III (public services, law enforcement, migration, justice) high-risk; Art. 27 (FRIA for public authorities).
Read next
Recruitment chatbots: transparency duty and high-risk once they select
A recruitment chatbot must always disclose that it is AI (Art. 50). Once it pre-selects, scores or rejects candidates, it is also a high-risk system (Annex III), with human oversight and the GDPR on top. A "handy assistant" thus quickly becomes a decision tool.
The algorithm register: must governments publish their AI?
Dutch public bodies publish the algorithms they use in the national Algorithm Register, as a transparency instrument. In addition, the AI Act requires registration of high-risk AI in an EU database (Art. 49/71) — also for public authorities as deployers. Two registers, one aim: accountability.
AI for dynamic freight pricing: what does the AI Act say?
AI that sets freight prices dynamically is not as such on the AI Act's high-risk list. AI literacy and transparency duties do apply, and the classification can shift where there are effects on individuals.