Skip to content
For insurance

Insurance pricing AI is high-risk by name, not by interpretation.

Most industries have to interpret whether Annex III covers them. Insurance doesn't. The Act explicitly lists AI used for "risk assessment and pricing in relation to natural persons in the case of life and health insurance." If that's your product, you have until December 2, 2027 to comply.
01 / High-risk

Which insurance AI is high-risk?

Annex III, point 5 covers AI affecting access to essential private services, and names life and health insurance specifically.

Life & health risk assessment

AI that evaluates individual risk profiles for life or health insurance, named directly in Annex III, point 5(c). This isn't an interpretation; it's in the text.

Premium pricing for life & health

Systems that calculate or influence individual premiums for life and health policies, including telematics-style health scoring and wearable data models.

Underwriting eligibility decisions

AI that decides or materially influences whether an applicant is offered life or health coverage, including automated decline rules.

Creditworthiness in premium financing

If you assess creditworthiness for payment plans or premium financing, that's a separate Annex III trigger under point 5(b).

02 / Probably fine

What's not high-risk in insurance?

The high-risk listing is narrower than most insurers assume: it covers life and health lines, not the whole industry.

  • Property & casualty pricing (5(c) names only life and health insurance)
  • Motor insurance telematics for vehicle pricing (unless tied to health assessment)
  • Internal actuarial modelling and reserving with no individual-level decisions
  • Fraud analytics that flag claims for human review without restricting access
  • Customer service chatbots (limited risk, Article 50 transparency applies instead)

Even if your system isn't high-risk, Article 50 transparency may still apply, especially for customer-facing chatbots and AI-generated communications.

03 / Obligations

What high-risk insurance AI must do

Each must be in place before December 2, 2027. Non-compliance risks fines up to €15 million or 3% of global turnover.

  1. 01Risk management system (Article 9)
  2. 02Data governance & bias documentation (Article 10)
  3. 03Full Annex IV technical documentation
  4. 04Automatic event logging (Article 12)
  5. 05Transparency & instructions for deployers (Article 13)
  6. 06Human oversight measures (Article 14)
  7. 07Accuracy, robustness & cybersecurity (Article 15)
  8. 08Conformity assessment (Article 43)
  9. 09EU database registration (Article 49)
  10. 10Post-market monitoring (Article 72)
04 / Overlap

Already regulated under Solvency II and GDPR?

Insurers are used to heavy regulation, and existing model governance under Solvency II plus GDPR DPIAs will partially cover the AI Act's risk management (Article 9) and data governance (Article 10). But the AI Act adds what your actuarial governance doesn't touch: bias documentation across protected groups, conformity assessment, EU database registration, automatic logging, and instructions for deployers. EIOPA has signalled AI Act supervision will run alongside, not replace, existing insurance supervision.

See the full GDPR overlap mapping

Find out where your insurance AI stands.

Insurance AI is named in the regulation itself, so supervisors won't debate whether you're in scope. The classifier is free; classify your system and start generating the documentation you need.

Classify your AI system