# Texas Bulletin B-0003-26 vs the NAIC model bulletin: every difference, with the evidence on both sides

> Texas is recorded as 'Independent framework' on the NAIC adoption map, under Bulletin B-0036-20. Texas has two instruments and the NAIC map points at the older one. B-0036-20 (2020) is about the accuracy of third-party data; the AI bulletin is B-0003-26, issued June 2026. Its opening paragraph tracks the model bulletin almost word for word, then goes its own way: it names the NAIC's 2020 AI Principles as its guide, not the 2023 model text.

- Source page: https://insureaiwire.com/states/texas/
- Publication: InsureAI Wire
- Difference layer reviewed: 2026-08-18
- Measured against: naic-model-bulletin-2023-12-04 (parallel-framework)
- NAIC AI Systems Evaluation Tool pilot: not participating. Its own examinations and product filings are what apply instead. (https://content.naic.org/sites/default/files/call_materials/Pilot%20Project%20Summary.pdf)
- Method: every departure below is quoted from both instruments with a locator. Where one side has nothing to quote, the reading that established the absence is given in its place. Not legal advice.

## What is different

| Point | NAIC model | Texas |
|---|---|---|
| Human review | — | A person must approve consequential decisions |
| AI in utilisation review | — | Statute bans AI adverse determinations |
| Written AI programme | A written programme of record | — |
| Vendors and outside data | A section of its own | One clause, no contract terms |
| Who it addresses | Insurers holding a certificate | Every regulated entity, agents and adjusters too |
| At examination | Expect to be asked | Furnish on request; four named areas |

An em dash means that instrument says nothing on the point. That is a finding, not a gap in the review.

## Every difference, with both texts

### Who and what it covers

#### Covers more insurers: Texas addresses every regulated entity, plus their agents and representatives. The model bulletin speaks to insurers holding a certificate of authority and to nobody else.

- Item id: tx-scope-regulated-entities
- NAIC model bulletin (Section 1, p.1): 'to remind all Insurers that hold certificates of authority to do business in the state that decisions or actions impacting consumers that are made or supported by advanced analytical and computational technologies, including Artificial Intelligence (AI) Systems … must comply with all applicable insurance laws and regulations.'
- Texas (B-0003-26, header): 'To: All regulated entities and their agents and representatives'
- Supporting reading: The model bulletin never addresses the distribution chain: "agent" 0 hits and "adjuster" 0 hits (word-boundary counts, pypdf, 2026-08-18). Texas cites the agent and adjuster licensing chapters (4001 and 4101) among the laws that apply.
- Source: Bulletin B-0003-26 · Texas TDI, retrieved 2026-08-18 via curl (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)

### The written programme

#### Extra requirement here: Texas puts a person in the loop by rule. Where AI makes a consequential decision, someone has to review it and agree with it before anything happens to the customer.

- Item id: tx-human-review-consequential
- Not in the model bulletin. Basis for that reading: The model bulletin never requires human review: "consequential" 0 hits and the phrase "review and agree" 0 hits (word-boundary counts over the 9-page, 28,577-character text, pypdf, 2026-08-18). It treats "the extent to which humans are involved in the final decision-making process" as one of five factors that set how strong controls should be (Section 3 introduction, pp.4-5).
- Texas (B-0003-26, Guidance and Expectations): 'If a regulated entity uses AI to make a consequential decision, TDI expects a person to review and agree with all decisions before action is taken.'
- Source: Bulletin B-0003-26 · Texas TDI, retrieved 2026-08-18 via curl (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)

#### Not required here: The centrepiece of the model bulletin is absent. Texas asks for controls, governance and testing in general terms, but never for a written programme of record.

- Item id: tx-no-written-programme
- NAIC model bulletin (Section 3, p.4): 'all Insurers authorized to do business in this state are expected to develop, implement, and maintain a written program (an "AIS Program") for the responsible use of AI Systems that make, or support decisions related to regulated insurance practices.'
- Not in Texas. Basis for that reading: The word "program" does not appear anywhere in B-0003-26 (0 hits across 7,501 characters of body text, word-boundary regex, 2026-08-18), nor do "inventory" (0) or "attestation" (0). Texas closes with the model's own disclaimer instead: its goal "is not to prescribe specific practices or documentation requirements".
- Source: Bulletin B-0003-26 · Texas TDI, retrieved 2026-08-18 via curl (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)

### Vendors and outside data

#### Asks for less here: Where the model devotes a section to third-party data and systems, Texas extends its expectations to third parties in a single clause and leaves the mechanics to you.

- Item id: tx-third-party-one-line
- NAIC model bulletin (§4.1, p.7): 'Due diligence and the methods employed by the Insurer to assess the third party and its data or AI Systems acquired from the third party to ensure that decisions made or supported from such AI Systems that could lead to Adverse Consumer Outcomes will meet the legal standards imposed on the Insurer itself.'
- Texas (B-0003-26, opening): 'These expectations extend to any third party working with a regulated entity and are designed to protect Texans' privacy and interests, ensure freedom from unfair discrimination, and promote the knowledge and development of ethical AI systems.'
- Supporting reading: "third party" appears once in B-0003-26, in the clause quoted here (word-boundary count over 7,501 characters, 2026-08-18). There are no contract terms, no audit rights and no cooperation clause, all of which the model sets out at §4.2.
- Source: Bulletin B-0003-26 · Texas TDI, retrieved 2026-08-18 via curl (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)

### What an examiner can ask for

#### Same thing, different words: Both texts land in the same place at examination: produce what you have when asked. Texas names the four areas its questions will cover.

- Item id: tx-exam-furnish-on-request
- NAIC model bulletin (SECTION 4, p.7): 'Regardless of the existence or scope of a written AIS Program, in the context of an investigation or market conduct action, an Insurer can expect to be asked about its development, deployment, and use of AI Systems, or any specific Predictive Model, AI System or application and its outcomes'
- Texas (B-0003-26, Guidance and Expectations): 'All regulated entities using AI should be able to furnish procedures and protections upon request by TDI. Regulated entities should expect TDI's monitoring of AI procedures and protections to include inquiries concerning governance frameworks, risk management, data and privacy protections, and internal controls.'
- Source: Bulletin B-0003-26 · Texas TDI, retrieved 2026-08-18 via curl (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)

### The law it rests on

#### Extra requirement here: One Texas statute bans the use case outright: a utilisation review agent may not use AI to make an adverse determination. The bulletin lists it among the laws that already bind you.

- Item id: tx-utilization-review-ban
- Not in the model bulletin. Basis for that reading: The model bulletin bans no use of the technology anywhere: "utilization review" 0 hits, "adverse determination" 0 hits, and "prohibit" appears only in its recital of unfair trade practice law. pypdf, 2026-08-18.
- Texas (B-0003-26, Legislative Authority): 'Utilization Review Agents: Texas Insurance Code Chapter 4201, concerning Utilization Review Agents, prohibits using AI to make an adverse determination.'
- Lines affected: health
- Source: Bulletin B-0003-26 · Texas TDI, retrieved 2026-08-18 via curl (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)


## What the regulator asks for

| Obligation | Who | Cadence | Due | Citation |
|---|---|---|---|---|
| Be able to hand over your AI procedures and protections when TDI asks. Questions will cover governance, risk management, data and privacy, and internal controls. | Any regulated entity using AI | on-request |  | B-0003-26, Guidance and Expectations |
| Product filings are one of the two channels TDI says it will use to watch AI, alongside examinations. What you put in a filing is where the questions start. | Any regulated entity filing products in Texas | continuous |  | B-0003-26, Guidance and Expectations |

- B-0003-26, Guidance and Expectations: 'All regulated entities using AI should be able to furnish procedures and protections upon request by TDI. Regulated entities should expect TDI's monitoring of AI procedures and protections to include inquiries concerning governance frameworks, risk management, data and privacy protections, and internal controls.'
- B-0003-26, Guidance and Expectations: 'TDI will monitor the use of AI through examinations and product filings. TDI also welcomes any complaints from consumers regarding the use of AI and will investigate accordingly.'

NAIC AI Systems Evaluation Tool pilot: not participating. Its own examinations and product filings are what apply instead. (https://content.naic.org/sites/default/files/call_materials/Pilot%20Project%20Summary.pdf)

## If you write in more than one state

Texas is the state where the obligation is a decision rule rather than a document. Nothing is filed and no programme is demanded, but a person has to review and agree with every consequential AI decision before it reaches the customer. That is a change to how the process runs, not to what is written down about it. It also reaches further down the chain than any other state here: agents and adjusters are addressed directly, so a carrier's answer has to cover the people selling and adjusting as well as the models. (This paragraph is InsureAI Wire's reading, resting on: tx-human-review-consequential, tx-scope-regulated-entities, tx-utilization-review-ban, tx-furnish-procedures.)

## What InsureAI Wire has written about Texas

- Texas TDI Bulletin B-0003-26 Sets AI Governance Expectations for Regulated Entities (https://insureaiwire.com/news/texas-tdi-bulletin-b-0003-26-ai-governance/) · Bulletin B-0003-26, June 2026: what TDI now expects of every regulated entity using AI in consumer-facing decisions.
- Texas Bulletin B-0036-20 on Third-Party Data (https://insureaiwire.com/news/texas-tdi-b-0036-20/) · The 2020 bulletin that makes the carrier, not the vendor, answerable for third-party data in rating, underwriting, and claims.
- Texas Enacts AI Governance Laws for Healthcare (https://insureaiwire.com/news/texas-hb-149-sb-1188-ai/) · HB 149 and SB 1188, including the subsection that carves insurers back out of the discrimination provisions.

## Sources read

- Commissioner's Bulletin B-0003-26, Use of artificial intelligence, June 12, 2026 · Texas TDI, retrieved 2026-08-18 via curl (HTML, 38,288 bytes; 7,501 characters of body text after stripping markup) (https://www.tdi.texas.gov/bulletins/2026/B-0003-26.html)
- Commissioner's Bulletin B-0036-20, Insurers' use of third-party data, September 30, 2020 · Texas TDI, retrieved 2026-08-18 via curl (HTML, 29,506 bytes) (https://www.tdi.texas.gov/bulletins/2020/B-0036-20.html)
- Model Bulletin: Use of Artificial Intelligence Systems by Insurers, adopted December 4, 2023 · NAIC, retrieved 2026-08-18 via curl + pypdf (9 pages, 28,577 characters) (https://content.naic.org/sites/default/files/inline-files/2023-12-4%20Model%20Bulletin_Adopted_0.pdf)
- AI Systems Evaluation Tool Pilot: Pilot Project Summary · NAIC, retrieved 2026-08-18 via curl + pypdf (357,714 bytes; the 12-state list on p.1 does not include Texas) (https://content.naic.org/sites/default/files/call_materials/Pilot%20Project%20Summary.pdf)

Last reviewed Aug 18, 2026.