At a glance
| Jurisdiction tier | Tier C - Specified, not yet enforced. AI is specified as for Tier A, but no mechanism to check firms against it is stated. |
| Case examined | Meiji Yasuda Life - an AI risk-prediction model in life underwriting |
| Band for this case | Elevated |
| Rules reach | Insurers: directly · Reinsurers: directly · Intermediaries: directly |
| Last verified | [actual date of the launch-window check] |
Japan's AI Act was promulgated and took effect on 4 June 2025; its chapters on the AI Basic Plan and the AI Strategy Headquarters took effect on 1 September 2025. It requires businesses only to strive to use AI and to cooperate with government measures, and it carries no penalties. This edition found no AI-specific insurance rule.
The Financial Services Agency's AI Discussion Paper (version 1.1, March 2026) says it sets no supervisory expectations. It also says that where AI is used in a significant decision, the institution itself needs to be able to verify and explain the reasonableness of that decision.
Who these rules reach
- Insurers: directly, under the AI Act's Article 7, which applies to "活用事業者" (businesses that use AI); its duties are to strive and to cooperate.
- Reinsurers: directly, under the same Article, as businesses that use AI.
- Intermediaries: directly, under the same Article, as businesses that use AI.
This band describes how clearly the rules meet the AI use that Meiji Yasuda Life has itself published. It is not a finding that Meiji Yasuda Life has breached any rule.
Why Meiji Yasuda Life: Meiji Yasuda's own release of 21 January 2025 says it has introduced an AI predictive model into its life-insurance underwriting.
The question
In briefMeiji Yasuda Life announced on 21 January 2025 that it had introduced a predictive model using artificial intelligence into its life-insurance underwriting for a cardiovascular-disease product sold from 6 January 2025. The model is the company's own risk prediction, combined with conventional medical underwriting, and the company said the combination lets it accept more customers. The Financial Services Agency's AI Discussion Paper (version 1.1, March 2026) says it sets no supervisory expectations. It also says that where AI is used in a decision that is significant given its impact (credit decisions are its example), the institution itself needs to be able to verify and explain the reasonableness of that decision. Where the model's prediction bears on whether an applicant is accepted, does Meiji Yasuda count that among the significant AI-assisted decisions the FSA describes, able to verify and explain the model's part in each outcome? Or does it regard the medical underwriter's assessment as the decision, with the model as one input?
Why this band: The question points to one firm, one model and one boundary, but the regulator's statement it rests on is a discussion paper that sets no requirements.Meiji Yasuda says it combines an AI risk-prediction model with medical underwriting to assess life-insurance applicants, and that this lets it accept more of them. The FSA has said, without making it a requirement, that a firm using AI in a significant decision should be able to verify and explain that decision. Meiji Yasuda's announcement does not say how it would explain the model's part in an individual applicant's outcome.
3. The same question for other firms in Japan
Any insurer in Japan that uses an AI model in deciding whether to accept an applicant faces the same point: whether it treats that as a significant AI-assisted decision it can verify and explain case by case. This entry gives no band to any firm other than Meiji Yasuda Life.
4. What this entry does not establish
- Whether Meiji Yasuda used the model in 2026. No 2026 publication was found saying so; a page captured on 28 September 2026 shows only that the product and its services are live.
- Whether any applicant is declined or rated because of the model, whether an underwriter reviews every case, or what applicants are told. This entry says nothing either way.
- The FSA's list of licensed insurers was not read; the company is identified from its own release and the FSA's August 2026 list of internationally active insurance groups.
- Launch-window check: any Meiji Yasuda publication confirming or withdrawing the model.
- By 17 July 2028: commencement of Japan's amendment to its personal information protection act.
Sources
- 明治安田生命保険相互会社, 「生命保険の引受査定にＡＩを活用したリスク予測を導入！」, 2025年1月21日 (printed). Company's own, Japanese. meijiyasuda.co.jp PDF
- 明治安田, 「循環器病 対策Ｐｒｏ サービス」 page (undated; captured 28 September 2026). Company's own, Japanese. meijiyasuda.co.jp
- Financial Services Agency, AIディスカッションペーパー（第1.1版）, March 2026 (printed); published 3 March 2026. Regulator, Japanese. fsa.go.jp PDF
- Financial Services Agency, Publication of the summary of "Annual Report on Insurance Monitoring 2026", August 6, 2026 (printed). Regulator. fsa.go.jp
- 人工知能関連技術の研究開発及び活用の推進に関する法律 (令和七年法律第五十三号; Act No. 53 of 2025, the AI Act), promulgated 4 June 2025, Article 7 and Supplementary Provisions Article 1. Statute, Japanese. laws.e-gov.go.jp PDF
- 個人情報保護委員会 (Personal Information Protection Commission), release of 17 July 2026 on the promulgation of the amendment to the Act on the Protection of Personal Information. Regulator, Japanese. ppc.go.jp
None share a law or regulator with this entry.
________________________________________
Commercial relationship with any party named here: None · Right of reply: indexreport.protocol@fijishi.com; with the sender's agreement, replies are published in full below the entry, dated · Cite as: Fijishi Jurisdiction Index 2026, Entry JP · Corrections: none.