• Atlanta
• Boston
• Dallas
• Ft. Lauderdale
• London
• Minneapolis
• New York
• Oakland
• Philadelphia
• Washington, DC

by Jennifer Gibbs

On June 12, 2026, the Texas Department of Insurance (TDI) issued Commissioner's Bulletin No. B-0003-26, addressing the use of AI (and other advanced analytical and computational technologies) by regulated entities and their agents and representatives. The bulletin reminds industry participants that any decision or action affecting consumers, whether made or merely supported by AI, must comply with all applicable insurance laws and regulations, including those governing unfair trade practices and unfair discrimination.

Purpose and Scope

TDI's stated goal is consumer protection. The bulletin sets expectations for how regulated entities should govern the development, acquisition, and use of AI across their operations, and these expectations extend to any third party working with a regulated entity. TDI points to the NAIC's 2020 Principles on Artificial Intelligence and the Texas Department of Information Resources' AI Code of Ethics and Minimum Standards as useful frameworks, noting that the latter promotes human oversight, fairness, accuracy, redress, transparency, data privacy, security, and accountability.

Applicable Legal Framework

The bulletin catalogs the existing statutory provisions that TDI expects regulated entities to keep in mind when deploying AI, including Texas Insurance Code Chapters 541 (Unfair Methods of Competition and Unfair or Deceptive Acts or Practices), 542 (Processing and Settlement of Claims), 544 (Prohibited Discrimination), 560 (Prohibited Rates), 831 (Corporate Governance Annual Disclosure), 4001 (Agent Licensing), 4101 (Insurance Adjusters), 4201 (Utilization Review Agents, which expressly bars using AI to make an adverse determination), 751 (Market Conduct Surveillance), and 401 (Audits and Examinations). Notably, TDI is not creating new obligations but confirming that these existing laws apply regardless of the technology used.

Key Expectations

TDI's guidance emphasizes several core expectations:

• AI-driven decisions must not be inaccurate, arbitrary, capricious, or unfairly discriminatory, and compliance with these standards is required irrespective of the tools used.
• Regulated entities should implement controls and guardrails to mitigate the risk of adverse consumer outcomes.
• For consequential decisions made using AI, a human must review and agree with the decision before any action is taken.
• Strong governance, risk management, and internal audit functions—along with verification and testing methods to catch errors and bias—are viewed as central to compliance.
• TDI will monitor AI use through examinations and product filings, will investigate consumer complaints about AI, and expects entities to be able to produce their AI governance procedures and protections upon request.

Practical Takeaway

The bulletin does not prescribe specific practices or documentation formats. Instead, it puts regulated entities on notice that TDI examinations and market conduct reviews may probe AI governance frameworks, risk management processes, data privacy protections, and internal controls, including questions about specific AI applications. Entities using AI in underwriting, claims, rating, or other consumer-facing functions should be prepared to demonstrate human oversight and documented controls.

What This Means for the Industry

In plain terms, the bulletin states that an insurer cannot point to an algorithm as the reason a decision was fair, accurate, or lawful. Existing legal standards still apply in full, and insurers must now be prepared to demonstrate how those standards were met. For companies already using AI in underwriting, pricing, claims triage, or customer service, this bulletin is a signal to take a hard look at whether a human is genuinely reviewing and signing off on significant decisions, not just rubber-stamping outputs the system already produced. It is also a reminder that vendor tools count too. If a third-party AI platform is doing the heavy lifting, the company using it remains accountable if the outcome discriminates unfairly or cannot be explained.

Practically, this means maintaining a governance file documenting what the AI tool does, how it was tested for bias and accuracy, who reviews its outputs, and how errors are corrected if they occur. Entities that already maintain model governance programs in other contexts, such as credit or fraud models, have a head start, but should extend those programs specifically to insurance decisions such as underwriting and claims handling. 

_________________________________

The opinions expressed are those of the authors and do not necessarily reflect the views of the firm or its clients. This article is for general information purposes and is not intended to be and should not be taken as legal advice.