The experts who ask permission have already solved this.

There is a new listener in some examination rooms, and I think it belongs there. Ambient AI scribes are now in routine use across Australian clinical practice. The software listens to the consultation and a draft note appears at the end of it. Some of those same doctors also do medicolegal work, and some are bringing the tool with them.

That's a good thing. Documentation is the single heaviest tax on expert time in this industry, and it is the reason capable specialists quit medicolegal work. What interests me is that a handful of experts have already worked out how to do it properly. They ask permission at the start of the examination. They explain what the tool does. And they check the draft against their own recollection before it becomes the record. That is the whole answer. It just isn't standard practice yet.

A clinical note and a medicolegal record are different objects

A consultation note exists to help the treating team. It can be brief, it can summarise, and it can leave things out, because the reader is a clinician who wants the salient points. Examination notes in a medicolegal matter do different work. They can be sought, compared against the final report, and put to the expert under cross-examination. Their value comes from being a contemporaneous account of what happened, not a useful summary of it.

Ambient scribes are built for the first job. Summarising is the feature, not a defect. In a clinical setting, dropping detail is what makes the note readable. In an evidentiary setting, dropped detail is a gap someone will eventually ask about. Same tool, different standard, and the expert is the one who has to bridge it.

Know what the tool is good at

A scribe captures what is said. A medicolegal examination turns substantially on what is observed and measured. Range of motion. Gait. Guarding. Consistency of effort across the assessment. Behaviour during the history that never gets spoken aloud by anyone. None of that reaches the microphone unless the expert narrates it. So the note becomes rich in the conversational parts of the examination and thin in exactly the parts a report is most often challenged on.

That is not an argument against the tool. It's an argument for knowing what it does. The second thing to know is that these systems produce fluent prose, and fluency is not accuracy. A summary can tidy a hesitant answer into a clean one, or read a topic as addressed when the exchange was ambiguous. The expert who reviews the draft against their own recollection, while the examination is still fresh, is what turns a draft into a record.

Three records, two gaps

What was said and done. The examination itself. Unrecoverable once the room empties, unless something captured it. What was captured. The contemporaneous note, traditionally the expert's own. What was reported. The final report, which is what the parties actually receive and rely on.

Two gaps sit between those three records, and they have always existed. What changes with a scribe is who is bridging the first gap. Software now stands between what happened and what was captured, and the person relying on the report has not been told it is there.

What the experts already doing this get right

The experts I know who use scribes in medicolegal work do three things, and none of them are complicated. They ask. Permission is sought at the start, the examinee is told what the tool does, and the request is recorded. They verify. The draft is checked against their own recollection before it becomes the record, and the observed and measured findings are their own work rather than the software's. They would say so if asked. Not buried in a methodology paragraph nobody reads. Available, on the record, testable like any other part of their method.

An examinee in an IME is not the expert's patient in the usual sense, and recording and consent rules differ across Australian jurisdictions, which is a matter for proper advice rather than a newsletter. Asking clearly and noting the answer removes most of the difficulty before it starts.

Where the framework lands

Augmentation. Transparency. Verification. Augmentation is exactly what a scribe is: the expert's attention goes to the examination while the software handles the typing. Verification is the expert checking the draft against what they remember, because the record has to be theirs. And Transparency is the part still missing at an industry level, because a method that can be disclosed can be tested, and a method that can be tested is worth more than one taken on trust.

The bottom line

Scribes are in the examination room and they should be. They give an expert back the one thing this industry cannot manufacture, which is attention during the examination. The work left to do isn't prohibition and it isn't a policy nobody reads. Ask permission, verify the draft, be willing to say so. The experts already doing it have shown it takes about thirty seconds at the start of an examination.

If you instruct experts: would you want to know an AI scribe was used, and would it change how you read the report? And if you already ask permission before using one: what do you actually say, and how do examinees respond? That answer is more useful to this industry than anything I could write.

Next issue: every referrer says the industry needs more experts. Almost nobody wants to be the one who briefs an untested one. What it actually costs the person who puts their credibility behind a new expert first.