Artificial intelligence is reshaping legal practice faster than the law can keep up. A case now before the Texas Supreme Court — In re Patrick Hughey, set for oral argument in October 2026, sits at the center of that collision. At stake is whether AI-assisted transcription tools can legally replace certified court reporters in Texas depositions. The answer will send a clear signal about how courts will treat AI-generated work product in litigation for years to come.
The Issue
A wrongful termination case became a landmark technology dispute when the plaintiff's attorney used Skribe.ai, an AI-powered platform that records depositions via Zoom and generates transcripts using automated speech recognition (speech to text technology), reviewed by a human transcriptionist, instead of a traditional court reporter. The opposing party moved to strike the transcript, arguing Texas law requires a licensed court reporter to produce any written deposition transcript. The trial court agreed and struck the transcript.
The question before the Supreme Court: Does existing Texas law prohibit AI-assisted transcription tools from producing legally usable deposition records? The parties urge the court to define non-stenographic reporting with more clarity as it relates to evolving speech to text technology and AI.
What We're Seeing
This case is the first major Texas test of how AI-generated transcripts will be treated by the courts. Similar objections to AI-assisted transcripts have already surfaced in at least two other Texas cases, signaling a growing conflict between rapidly advancing technology and the existing legal framework governing deposition practice. The dispute centers on whether the Texas Rules of Civil Procedure, which expressly permit non-stenographic depositions and "written transcriptions" derived from them, allow AI tools to fill the role traditionally held by licensed court reporters. A critical 2023 legislative amendment replaced "recorded" with "reported" in the governing statute and directed that it be construed consistently with the rules permitting non-stenographic alternatives, a change that Relator argues explicitly opens the door to AI-assisted transcription. Opponents counter that no statute or rule authorizes an AI platform to produce an official legal record, and that doing so creates accountability gaps no technology can fill.
Key Takeaways for Clients
The Court's ruling will be one of the first in the country to directly address whether AI-generated transcripts constitute legally valid deposition records, establishing precedent well beyond Texas.
If AI-assisted transcription is permitted, the cost savings are significant, estimated at 45% or more compared to traditional court reporting, meaningfully improving access to justice in lower-value cases.
If prohibited, existing AI-generated transcripts across Texas litigation could be vulnerable to challenge, creating immediate risk for matters already in progress.
The decision will shape how firms evaluate AI-powered legal tools more broadly. Courts that reject AI transcription today may apply similar scrutiny to AI-drafted documents, AI-reviewed contracts, and other automated legal work products tomorrow.
The growing national shortage of certified shorthand reporters, down 21% since 2015, means that if AI tools are restricted, scheduling delays and increased costs will follow, particularly in rural jurisdictions.
Bottom Line
In re Patrick Hughey is not just a discovery dispute, it is an early-stage reckoning with how the legal system will govern AI-generated transcripts. The Texas Supreme Court's answer will reverberate far beyond deposition practice, and every firm advising clients on legal technology adoption should be watching closely.
The Case
In re Hughey, ___ S.W.3d ___, 2025 WL 1523269 (Tex. App.—Beaumont May 29, 2025), argument granted on pet. for writ of mandamus (May 29, 2026) [25-0463]

/Passle/MediaLibrary/Images/2026-02-13-21-18-43-901-698f953399773b21433b3f7f.png)
/Passle/6932ed9acb3df5dbeef98cc4/SearchServiceImages/2026-05-25-20-41-20-115-6a14b3f07706964e8615fde2.jpg)
/Passle/6932ed9acb3df5dbeef98cc4/SearchServiceImages/2026-05-18-16-58-34-407-6a0b453a52418ebe6168f24b.jpg)