Because artificial intelligence excels at pattern matching, the only irreplaceable value of a human court reporter is the real-time phonetic resolution of novel, unpredictable speech.

For fifteen years, the court reporting profession argued about stroke counts. Brief-heavy or phonetic. Write short or write through. Then AI arrived, and the debate stopped being about pedagogy.

It became about survival.

In a real deposition, speech is hostile. Witnesses invent terms. Lawyers quote obscure statutes. Names appear with no brief. Experts use technical vocabulary from fields the reporter has never encountered. Speakers overlap. Dialects vary.

The brief-trained reporter hits a wall. The brief doesn’t exist, and the phonetic foundation required to write through the novel word was not strong or was never built.

Brief-heavy methods are highly effective in structured, predictable testing environments — which is exactly what certification exams are, and exactly what real proceedings are not.

When a language model encounters input it cannot resolve through pattern matching, it does the same thing: it guesses. It hallucinates. In a legal transcript, a hallucination is not a technical glitch. It is testimony that was never given, attributed to a witness who never said it, in a document that may be introduced as evidence.

The phonetically grounded reporter does something the machine structurally cannot. They don’t guess. They write through it. They capture the sound, syllable by syllable, securing the verbatim record precisely when the predictable patterns fail.

Real-time phonetic resolution of novel speech is the profession’s only irreplaceable value.

Brief-heavy training didn’t just fail to solve the attrition crisis. It optimized the small number of students who survived it for the part of the job AI does best — and left them underprepared for the part AI cannot do.

Read that again.

The profession’s survival argument and its training argument are the same argument. We don’t need more people who can compete with machines at pattern recognition. We need professionals with the phonetic architecture to secure the record when the patterns break.

That is what a court reporter is. That is what training must produce.