AI governance and expert responsibility
How this platform is designed around expert independence, and safe-use principles for practitioners.
AI may assist with quality-control tasks, but responsibility for expert evidence remains with the expert. Expert Witness QA is designed around that principle.
The platform provides prompts for consideration. It does not require the expert to make any amendment and does not mandate any conclusion. Any change made to a report after using the platform should reflect the expert’s own independent judgement, reasoning and professional opinion.
Experts should verify any AI-assisted output against the underlying source material before relying on it. This includes the letter of instruction, medical records, examination findings, imaging reports, witness evidence, occupational information, care evidence, published literature and any other documents relied upon.
Safe-use principles
- Use the output as a checklist, not as an opinion.
- Check all flagged issues against the source documents.
- Do not adopt suggested changes unless you independently agree with them.
- Keep a proportionate record of meaningful AI-assisted QA use where appropriate.
- Consider whether the instructing party should be informed of AI-assisted quality-control use.
- Do not upload identifiable, privileged or confidential material unless appropriate safeguards and authority are in place.
- Do not use the platform to ask for diagnosis, causation, prognosis, liability or legal strategy.
Independence and Part 35
The expert’s duty to the court is personal. Use of this platform must not compromise independence or create the appearance that the expert’s opinion has been directed by a third party or by an AI system.
The platform is intended for report-quality checking and instruction-coverage review. Where a review identifies possible evidential gaps, caveats or questions for clarification, these are prompts for the expert’s independent consideration only. They should not be treated as advice on litigation strategy or as directions about the expert’s opinion.
Disclosure and record keeping
Whether AI-assisted QA use should be disclosed will depend on the circumstances, the extent of use, the nature of the material processed, the instructions, and any procedural or court expectations. Users should take legal or professional advice where appropriate.
Even where disclosure is not required, experts should be able to explain what the tool was used for and how any output was checked.