AI ethics insurance

Category: AI in insurance · Reviewed by Tim Roche, Director · PI & Commercial · Last reviewed 2026-06-10

AI ethics in insurance is the body of principles, governance and practice that addresses fairness, accountability, transparency, safety and contestability when artificial-intelligence and machine-learning systems are used in insurance. In the United Kingdom market it sits at the intersection of the Equality Act 2010, the UK GDPR, the FCA’s outcomes-based supervision (anchored in the Consumer Duty), the EIOPA AI Governance Principles, and the HM Government AI White Paper’s five cross-economy principles.

Category: AI in insurance · Aliases: AI fairness insurance, ethical AI, responsible AI · Established: EIOPA AI Governance Principles June 2021; HM Government AI White Paper March 2023; FCA Consumer Duty July 2023/2024 · Related: Explainable AI insurance, AI underwriting, Machine learning underwriting

Definition

AI ethics in insurance focuses on the substantive risks that AI systems pose to:

• fairness — direct, indirect and proxy discrimination on protected characteristics;
• accountability — clarity over who, in the firm, is responsible for an AI-driven outcome;
• transparency — meaningful disclosure to consumers and regulators;
• safety / robustness — model behaviour under stress and adversarial conditions; and
• contestability — the right and the practical ability of an affected person to challenge an AI-driven decision.

Legal / Regulatory basis

• Equality Act 2010 — direct and indirect discrimination on the basis of protected characteristics (age, disability, sex, race, religion, sexual orientation, gender reassignment, marriage / civil partnership, pregnancy / maternity). Pricing or claims decisions that indirectly disadvantage a protected group, without objective justification, can amount to indirect discrimination.
• UK GDPR — Articles 5 (fairness, accuracy, data minimisation), 9 (special category data), 22 (automated decisions) and 35 (DPIA).
• Data Protection Act 2018 — Schedule 1 conditions for special category and criminal data processing.
• FCA Handbook PRIN (Consumer Duty, PRIN 2A) — outcomes-based regulation: products and services, price and value, consumer understanding, and consumer support.
• FCA & PRA DP5/22 and FS2/23 — supervisory expectations on AI governance.
• HM Government, A pro-innovation approach to AI regulation, White Paper March 2023 — five cross-economy principles: (i) safety, security and robustness; (ii) appropriate transparency and explainability; (iii) fairness; (iv) accountability and governance; (v) contestability and redress. The February 2024 Response confirmed the approach of regulator-led implementation.
• EIOPA, Artificial Intelligence Governance Principles, June 2021 — six principles: proportionality, fairness and non-discrimination, transparency and explainability, human oversight, data governance and record-keeping, robustness and performance.
• IAIS draft Application Paper on the Supervision of AI (2024 consultation) — international supervisory benchmark.
• EU AI Act (Regulation (EU) 2024/1689) — for UK firms with EU operations.
• ICO and Alan Turing Institute Explaining decisions made with AI (May 2020); ICO generative AI work (2024).
• FCA Innovation Hub and AI Sandbox cohorts — supervisory engagement on fairness and ethics design.

How it works in practice

A UK insurance firm’s AI ethics programme typically includes:

1. Use-case inventory — every AI / ML model is registered with owner, tier, customer impact and protected-characteristic risk.
2. Fairness assessments — slicing performance and outcomes by protected characteristics or close proxies; documenting any disparity and the firm’s response.
3. Data minimisation — challenging whether each input variable is necessary, particularly any close proxy for protected characteristics.
4. DPIA under UK GDPR Article 35 for high-risk processing.
5. Human oversight — meaningful review of consequential decisions; documented escalation routes.
6. Contestability — the customer can request and obtain a human review; this is built into the customer journey and the privacy notice.
7. Consumer Duty linkage — fair value, consumer understanding and consumer support outcomes are reported through the Consumer Duty board report.
8. Senior accountability — typically the SMF-3 / SMF-16 / SMF-24 holders share responsibility for AI risk under SMCR.
9. External commitments — alignment with the HM Government five principles and EIOPA’s six principles.

Common variations / Subsequent developments

• Specific scrutiny of pricing — the FCA’s price-walking remedy (ICOBS 6B) sits alongside fairness expectations on AI models.
• Proxy-discrimination guidance — increasing supervisory and industry attention to indirect discrimination through close proxies.
• Generative AI ethics — additional considerations on hallucination, provenance, intellectual property and child safety.
• Climate fairness — questions emerging on the affordability and availability of flood and subsidence cover, and AI’s role in those decisions.
• EIOPA Consultation on the Opinion on AI Governance and Risk Management (February 2025) — likely to harden EU expectations and influence UK practice.

Example

A UK home insurer’s AI ethics committee reviews a proposed GBM upgrade. Fairness slicing reveals a small premium uplift on customers in postcodes with high indices of multiple deprivation, partly attributable to genuine claims experience and partly to a close proxy variable. The committee instructs the modelling team to remove the proxy, document the small predictive-power loss, and re-evaluate. The decision is recorded in the firm’s AI register, the model-risk file and the Consumer Duty board report. A documented procedure exists for customers to request human review of automated decisions, consistent with UK GDPR Article 22.

See also

• Explainable AI insurance
• AI underwriting (batch 12 expansion)
• Machine learning underwriting
• AI customer service insurance

References

• Equality Act 2010, https://www.legislation.gov.uk/ukpga/2010/15
• UK GDPR Articles 5, 9, 22, 35; Data Protection Act 2018, Schedule 1, https://www.legislation.gov.uk
• HM Government, A pro-innovation approach to AI regulation, White Paper, March 2023; Government Response, February 2024. https://www.gov.uk
• FCA, PS22/9 — Consumer Duty, July 2022.
• FCA & PRA, DP5/22 / FS2/23 — AI and Machine Learning, October 2022 / 2023.
• EIOPA, Artificial Intelligence Governance Principles, June 2021. https://www.eiopa.europa.eu
• EIOPA, Consultation Paper on the Opinion on AI Governance and Risk Management, February 2025.
• IAIS, Draft Application Paper on the Supervision of AI, 2024 consultation. https://www.iaisweb.org
• ICO and Alan Turing Institute, Explaining decisions made with AI, May 2020.
• Regulation (EU) 2024/1689 (EU AI Act).

This entry is part of the Apex Insurance Wiki. Last reviewed by Matt Bartlett on 2026-06-10. Next review: 2026-12-10.

Apex Insurance Brokers Limited. Authorised and regulated by the Financial Conduct Authority, FRN 724952. Registered in England and Wales, Companies House 07014570. This entry provides general information about UK insurance concepts and is not regulated advice. Consult your insurance broker on your specific position.