Why a Texas Supreme Court case raises important questions about the future of court reporting, and the continued importance of trained human expertise.
Artificial intelligence is transforming the legal profession. From legal research and document review to transcription and evidence management, technology is creating opportunities to work faster, more efficiently and with greater accessibility.
Court reporting is undergoing its own transformation.
But as technology becomes increasingly capable of converting speech into written text, an important question emerges: Is producing a transcript the same as creating a reliable, defensible legal record? A recent case before the Texas Supreme Court brings that question into sharp focus.
In In re Patrick Hughey, a dispute arose over the use of an AI-assisted transcription platform, Skribe.ai, to record depositions and produce written transcripts without a certified shorthand reporter.
The opposing party challenged the approach, relying on Texas legislation governing court reporting and the preparation of deposition transcripts. The trial court quashed the disputed depositions and struck the transcript at issue.
The matter reached the Texas Supreme Court, which heard oral arguments on October 6, 2026.
The legal dispute concerns how Texas's existing statutory and procedural requirements apply to newer transcription technologies. It is not simply a determination of whether AI can accurately transcribe testimony. Nevertheless, the broader questions resonate throughout the legal profession, including Canada.
What constitutes a reliable legal record? Who is responsible for its accuracy? And what qualifications should be required of the person overseeing its creation? These questions deserve careful consideration as the industry continues to evolve.
One of the most persistent misconceptions about court reporting is that the role is simply to capture words. Those of us who have spent our careers in this profession know it is considerably more complex.
Experienced court reporting professionals understand legal proceedings, terminology, speaker identification, procedural requirements and the importance of preserving testimony accurately and impartially. We recognize when an answer is unclear, when participants speak over one another, when a name or technical term requires clarification, or when something may compromise the integrity of the record. And importantly, we understand why those details matter.
Creating a legal transcript is fundamentally different from writing an essay or a novel. When we write, we control the words, sentences and structure. When we transcribe, we must accurately capture someone else's words, however fragmented, grammatically imperfect or difficult to understand they may be. Consider a witness whose first language is not English, or someone who speaks in incomplete thoughts, changes direction mid-sentence or uses unfamiliar terminology. A trained professional must draw on linguistic comprehension, context and experience to recognize what was actually said, without altering or interpreting the testimony to make it read more elegantly.
The objective is not to make the witness sound better. It is to preserve precisely what the witness said.
This is also why training competent legal court reporters and transcriptionists is far more challenging than simply teaching someone to use transcription software. Technical skills and formatting conventions can be learned, but the ability to recognize linguistic nuances, identify inconsistencies and know when something requires further investigation demands a different level of comprehension and judgment. Not everyone who can type accurately or write well possesses those skills. And the same challenge applies to AI: a transcript can appear perfectly coherent while containing subtle but consequential errors.
A transcript can read beautifully and still be wrong.
Consider the difference between: "I did authorize the transaction” and "I didn't authorize the transaction." A single word can fundamentally alter the meaning of testimony.
Technology can assist in producing transcripts with remarkable speed, but identifying errors, resolving ambiguities and ensuring that the final record accurately reflects the evidence require more than automated word recognition. It requires trained human judgment.
There is an important distinction between having a human review an AI-generated transcript and having a trained court reporting professional responsible for the integrity of the record.
A human reviewer may correct spelling, punctuation or obvious transcription errors. But without appropriate training, that individual may not recognize an error that appears entirely plausible on the page.
That is one of the more subtle risks of automated speech recognition. An incorrect word or phrase may be grammatically correct, make sense within the sentence and still be completely different from what the witness actually said.
A trained professional brings more than proofreading skills. They bring familiarity with legal language and procedure, an understanding of context and the judgment to recognize when something requires verification rather than assumption.
Equally important is knowing when not to make a determination.
A skilled professional understands that an unclear word cannot simply be replaced with one that seems logical. They know when to review the recording, investigate a possible error or acknowledge that something cannot be reliably determined.
Human oversight should therefore be more than a final quality-control step. It should be a meaningful part of the process, supported by training, defined standards and clear accountability.
The question isn't simply whether a human was involved in creating the transcript. It's whether that human had the expertise necessary to recognize errors and protect the integrity of the record.
None of this is an argument against artificial intelligence. Quite the opposite. Court reporting has always evolved alongside technology, from shorthand machines and computer-aided transcription, realtime reporting and now advanced speech recognition. Each innovation has changed how the work is performed and expanded what is possible.
AI presents substantial opportunities. Technology can accelerate transcript production, support realtime access to proceedings, improve searchability and help legal professionals work with the record more efficiently.
But the greatest opportunity lies in combining these capabilities with appropriately trained human expertise.
In a technology-assisted court reporting model, the professional's role may look different from that of a traditional stenographic reporter. Rather than personally capturing every word through shorthand, the professional may be responsible for monitoring the proceeding, ensuring recording quality, identifying speakers, documenting significant events and overseeing the transcription and verification process.
These are not incidental responsibilities. They are essential safeguards. And they require specialized training.
At Array, this philosophy informs our approach to technology-assisted court reporting. We recognize the enormous potential of technology to improve access to and engagement with the legal record, while maintaining the importance of trained professionals and meaningful human oversight.
The objective is not to replace professional judgment with technology. It is to use technology to support and enhance that judgment.
When lawyers order a transcript, they are not simply purchasing a document. They are relying on a record that may be used to prepare for trial, challenge credibility, support an appeal, resolve a dispute or establish precisely what a witness said under oath. They need confidence that the record is accurate, complete and capable of withstanding scrutiny. That confidence comes not only from the technology used to create it, but also from the processes, professional standards and people responsible for its production.
As AI becomes more sophisticated, the industry must continue to ask difficult questions about quality assurance, training, verification and accountability. Who monitors the proceeding when technology encounters difficulty? Who identifies and resolves discrepancies? Who determines when a recording is insufficient to support a reliable transcription? And ultimately, who stands behind the finished record?
These are not questions technology alone can answer.
The Texas case may ultimately help clarify the legal boundaries for AI-assisted deposition reporting in that jurisdiction. But regardless of the outcome, the more fundamental issue remains: how do we ensure the integrity of a legal record as the technologies used to create it continue to change?
The debate should not be reduced to whether traditional court reporting should give way to artificial intelligence. Nor should we assume that an established method is necessarily the only reliable or defensible method.
Instead, we should focus on the standards required to ensure that every legal record, regardless of how it is captured, is accurate, reliable and supported by appropriate professional accountability.
Different technologies may require different workflows and different skills. But the underlying obligation to preserve the integrity of the record does not change. As we embrace the opportunities AI presents, we must be equally committed to developing the trained professionals who will use it responsibly.
Because the future of court reporting isn't about choosing between people and technology. It's about ensuring that technology works in the service of professional expertise, and never at the expense of the record.
Further reading: AI in the Courtroom: The Texas Case That Could Redefine Court Reporting, published on JD Supra. Case reference: In re Patrick Hughey, Supreme Court of Texas, No. 25-0463.