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. |
| Question to | the Office of Insurance Commission - human oversight of AI in underwriting and claims |
| Band for this question | Contained |
| Rules reach | Insurers: directly · Reinsurers: not stated · Intermediaries: not stated |
| Last verified | [actual date of the launch-window check] |
The Office of Insurance Commission (OIC) published an AI Governance Guideline for insurers in November 2025. It classes AI in underwriting, premium-setting and claims as high-risk and says insurers should consider human oversight for those uses; it makes customer disclosure mandatory.
The Guideline says it links with OIC's related notifications and guidelines, and refers to OIC's enterprise risk management framework and its requirements on IT third-party risk.
Who these rules reach
- Insurers: directly, as the insurance companies the Guideline is written for ("สำหรับบริษัทประกันภัย", for insurance companies). The Guideline uses "ควร" (should) for its human-oversight expectation.
- Reinsurers: not stated.
- Intermediaries: not stated.
This band describes how far the published record answers the question put to the Office of Insurance Commission. It is not a finding that the Office of Insurance Commission, or any firm it oversees, has fallen short of any rule.
Why the question goes to the Office of Insurance Commission: The question concerns the regulator's own guideline, and this entry relies on no single company's account.
The question
In briefOIC's AI Governance Guideline for insurers, published in November 2025, classes AI used in underwriting, premium-setting and claims as high-risk and says insurers "should consider" human-in-the-loop or human-over-the-loop oversight for those uses - while saying customer-facing AI "must" be disclosed and must offer a switch to human staff. The Guideline says it links with OIC's related notifications and guidelines, and refers to OIC's enterprise risk management framework and its requirements on IT third-party risk. Does OIC treat the human-oversight expectation for high-risk underwriting and claims AI as enforceable through those notifications, or does it remain voluntary until OIC issues a notification of its own?
Why this band: No company's own account could be tested, so the question rests on the regulator's own published guideline.Thailand's insurance regulator has told insurers that AI in underwriting and claims is high-risk and should have human oversight - but in the language of a guideline, not a rule, while making only customer disclosure mandatory; OIC has not publicly stated whether that oversight expectation can be enforced through its existing notifications.
3. What the open question means for firms in Thailand
Until OIC answers, a Thai insurer using AI in underwriting or claims has no published statement on whether the human-oversight expectation can be enforced through OIC's existing binding notifications. This entry gives no band to any firm.
4. What this entry does not establish
- The notifications and guidelines the Guideline refers to were not read; the Guideline's own references to them are used.
- Whether any Thai insurer uses AI in these decisions. This entry says nothing either way.
- No dated item is recorded for this entry.
Sources
- Office of Insurance Commission, แนวปฏิบัติ เรื่อง การกำกับดูแลการใช้งานปัญญาประดิษฐ์ สำหรับบริษัทประกันภัย, November 2025 (พฤศจิกายน พ.ศ. 2568, printed). Regulator, Thai. oic.or.th PDF
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 TH · Corrections: none.