Skip to main content
AI Act high-risk sectors

The AI Act's high-risk landscape. Made legible.

Annex III of the EU AI Act lists eight categories of high-risk AI use. If your firm uses AI in any of them, you're in scope. cosantio specialises in the sectors most exposed under Annex III §4 (employment) and §5 (essential services, including insurance and financial services). The Scorecard works for all eight categories.

8 Annex III categories2 in deep focus2 Dec 2027 standalone high-risk deadline
The Annex III taxonomy

Eight categories of high-risk AI.

Article 6(2) of the EU AI Act classifies the systems in Annex III as high-risk. The eight categories below are the full taxonomy. The two highlighted rows are where cosantio's key focus sectors sit: employment under §4, and insurance and financial services under §5.

§1
Biometrics
Remote biometric identification, biometric categorisation, emotion recognition
Future scope
§2
Critical infrastructure
AI in road traffic, water, gas, heating, electricity supply
Future scope
§3
Education & vocational training
Admissions AI, exam scoring, behaviour monitoring, course allocation
Future scope
§4
Employment & worker managementEmployment & HR
CV screening, hiring scoring, performance monitoring, task allocation
Key focus
§5
Access to essential servicesInsurance · Financial services
Creditworthiness §5(b), life and health insurance pricing §5(c), benefits eligibility, emergency triage
Key focus
§6
Law enforcement
Risk assessment of offenders, polygraph use, evidence reliability scoring
Future scope
§7
Migration, asylum & border control
Visa application scoring, asylum risk profiling, border-stop polygraphs
Future scope
§8
Administration of justice
AI assisting judicial decisions, alternative dispute resolution scoring
Future scope
Key focus areaFuture scopeSource: Regulation (EU) 2024/1689, Annex III
Key focus areas

Two categories. Three deep-dives.

Three focus areas, all within Annex III §4 (employment) and §5 (essential services): hiring & HR, insurance, and financial services. The fourth card covers the remaining six Annex III categories (§1, §2, §3, §6, §7, §8); deep-dives for those are on the roadmap. The Scorecard works across all eight.

Annex III §5(b)
CBI · DORA
Central Bank of Ireland

Financial services.

Credit unions, retail credit firms, payments and EMI, MiCA CASPs. AI used in creditworthiness assessment and credit scoring of natural persons.

In-scope AI systems
Loan underwriting & credit scoringFraud detection & transaction monitoringCollections triage & arrears scoring
Live obligations
  • AI literacy trainingArt. 4
  • Customer chatbot disclosureArt. 50
  • FRIA (when high-risk regime applies)Art. 27
12-15 min readRead the guide
Annex III §5(c)
EIOPA · CBI · POG
Insurance regulators

Insurance.

MGAs, brokers, and carriers operating in EU markets. AI used in life and health insurance pricing and underwriting decisions.

In-scope AI systems
Pricing engines & quote generationUnderwriting risk assessmentClaims triage & fraud detection
Live obligations
  • AI literacy trainingArt. 4
  • EIOPA Opinion alignment2024 Opinion
  • FRIA (when high-risk regime applies)Art. 27
12-15 min readRead the guide
Annex III §4
WRC · DPC
Workplace & privacy

Employment & HR.

Any SME using AI in hiring, performance management, or task allocation. Often overlooked, always in scope, regardless of company size.

In-scope AI systems
CV screening & candidate rankingVideo interview scoringPerformance review & task allocation
Live obligations
  • AI literacy trainingArt. 4
  • Worker information rightsArt. 26(7)
  • GDPR Art. 22 + AI Act overlapArt. 86
10-12 min readRead the guide
Annex III §1, §2, §3, §6, §7, §8
Various
Sector-specific

Other high-risk sectors.

Education, biometrics, critical infrastructure, law enforcement, migration, justice. The Scorecard covers every Annex III category; bespoke guidance is on the roadmap.

Future deep-dive coverage
BiometricsCritical infrastructureEducation & trainingLaw enforcementMigration & asylumAdministration of justice
In the interim
  • Take the Scorecard (works for every sector)Free
  • Talk to us about your specific use20 min
Roadmap · 2026 H2Get in touch
Common across every sector

Whatever sector you're in, the Scorecard works the same way.

Eight plain-English questions. A regulator-grade PDF. The same four obligations covered in every result. Sector context layers on top, it doesn't replace the core.

01

Risk classification under Article 6

Minimal, limited, high-risk, or prohibited. Mapped to your specific Annex III category if you're high-risk.

02

Four live obligations checked

Article 4 literacy, Article 5 prohibited practices, Article 50 transparency, Article 11 register.

03

Sector context overlay

If you're in a regulated sector, your obligations under EIOPA, the Central Bank, the DPC, or WRC are mapped alongside the Act.

04

Action plan, prioritised

Prioritised list of artefacts to produce. Ordered by regulatory urgency, with a credible weeks-to-readiness estimate.

Find out where your sector stands.

Eight questions. Three minutes. A regulator-grade PDF that scores your firm against every AI Act obligation already in force, with the sector context that applies to you.