Menu ▾
Start sealing, $349/mo Live demos For attorneys Standing orders Sanctions record Verify free Enterprise Pick what it does Why now Pick the problem Start where you stand The Drift Desk What is a standing order Sanctioned for AI citations Privilege and AI Detectors compared Authorship marks compared When the law changes The 124 laws The products Pricing The Regulayer Receipt HumanMark The free Sealer For insurers Support Open verifier About FAQ Legal

The sanctions record · U.S. District Court, Western District of Texas

Gonzalez v. Texas Taxpayers and Research Association

U.S. District Court for the Western District of Texas · sanctions of approximately $4,000

Page verified 5 August 2026. We could not open the order from this session, so everything below is recorded from contemporaneous reporting and official statements rather than quoted from the document. We say so rather than imply we read it.

A small sanction that belongs on the record for one reason: it is another instance of the sequence that drives outcomes everywhere on this page. Denied first. Admitted after. And the tool was not a chatbot, it was a citation generator.

What happened

  • The filingDefective citations in filings in Gonzalez v. Texas Taxpayers and Research Association, Western District of Texas.
  • The denialCounsel first denied using artificial intelligence.
  • The admissionCounsel then admitted using a citation generator.
  • The sanctionApproximately $4,000 in fees and costs.

The tool question it raises

“I did not use AI, I used a citation generator” is a distinction a lot of practitioners believe in, and no court on this record has accepted it. The clearest statement of where the line actually falls is in New South Wales’s Practice Note SC Gen 23, which carves out spelling and grammar checkers, transcription, translation, formatting, search engines and conventional legal research databases, and catches everything that generates substantive content.

A citation generator produces a citation. A citation is substantive content. That is the whole analysis.

Where it fits

It is the fifth instance of the same sequence on this record, and by far the cheapest:

  • ~$4,000Gonzalez. Denied, then admitted.
  • $2,500Fletcher. Evaded; the court said candour would have cost less.
  • ReprimandFivehouse. Called it clerical first.
  • Special costsHussein. Concealment held to be misleading the court.
  • Criminal contemptKo v. Li. Denied in open court, then confessed.

Same first move, five very different endings. The variable is how long the denial was maintained and how thoroughly it was dismantled.

What to do about it

Seven days free, then $349 a month.

Every denial on this record was made without a record to consult. Remove the guessing and the denial never happens. A signed record of which AI you used and what a person checked, sealed to the exact file you deliver, made on your own machine and checkable free by anyone you send it to.

Start the free weekOr verify a real record, free →

For a firm rather than a practitioner, the same engine runs across the estate: Regulayer for enterprise. See also the attorney certification and every jurisdiction we track.

Sources

  • Gonzalez v. Texas Taxpayers and Research Association, W.D. Tex. The order was not opened from this session.
  • Recorded from contemporaneous sanctions surveys. The figure is approximate as reported; we have not verified it against the order.

Related

Checked 5 August 2026. The order was not read at source. Treat the detail, including the figure, as reported.

Common questions

What happened in Gonzalez v. Texas Taxpayers?

Counsel filed material containing defective citations, first denied using artificial intelligence, and then admitted having used a citation generator. Sanctions of approximately $4,000 in fees and costs followed.

Does a citation generator count as AI?

For the purposes of these rules, what matters is whether the tool produced substantive content that was filed without verification. New South Wales Practice Note SC Gen 23 draws the line explicitly: spelling and grammar tools, transcription, translation, formatting and conventional legal research databases are outside it, while anything generating substantive content is inside.

Why does the initial denial matter so much in these cases?

Because courts have repeatedly priced it. Fletcher v. Experian recorded that candour would likely have produced a lesser sanction. Fivehouse v. Department of Defense treated an initial claim of clerical error as a candour problem. Hussein v. Canada treated concealment as misleading the court. Ko v. Li turned it into a criminal contempt proceeding.

Information, not legal advice, and not advice on the law of any jurisdiction. Every entry is verified against the issuing body’s own document; where a source is reporting rather than the document, we say so.