Now available: The Expert Record, the Standard for expert witnesses using AI. Where to buy

The Expert RecordThe standard for expert witnesses using AI

AI in the Courtroom · Expert witnesses and AI

Osteoarthritis Right Shoulder Review, 2026 CanLII 116891

In the Osteoarthritis Right Shoulder review, a Veterans Review and Appeal Board panel found that a nurse practitioner's medical opinion bore the hallmarks of AI drafting without review. Two of four cited articles named the wrong authors and journals. The panel gave the opinion no weight and granted the claim on other evidence.

Primary document readMedical opinion given no weight; entitlement granted on other evidence October 15, 2026

What happened

The Osteoarthritis Right Shoulder review, reported as 2026 CanLII 116891, concerned a veteran who served in the Canadian Armed Forces Regular Force from January 7, 1963 to March 9, 1995. He applied for a disability award on March 9, 2026. Veterans Affairs Canada denied entitlement the next day, and he asked the Veterans Review and Appeal Board to review that decision. Two of his claims were heard together on October 15, 2026.

For the review, the veteran's nurse practitioner filed a five-page medical opinion dated June 25, 2026. The opinion rested on four scientific articles, each listed with authors, a journal, and a PubMed or PubMed Central address. This is the same board, and the same kind of document, as the earlier Keratoconus Entitlement Review, where a different panel also questioned a medical opinion that appeared to be machine drafted.

What the tribunal did

The Panel checked the references. It recorded that the first article was attributed to the wrong authors and cited to the wrong journal; the Panel named the five actual authors and placed the article in the Annals of Work Exposures and Health. The second article was also attributed to the wrong authors, and the Panel identified the Orthopaedic Journal of Sports Medicine as the correct journal. The Panel also found an internal error. The opinion said the condition was diagnosed well within 25 years of the veteran's release from service. On the opinion's own dates, the Panel wrote, the gap was more than 31 years.

From those errors the Panel drew a conclusion about how the document was made. It found the report not credible and said it showed "the hallmarks of an opinion that was written by Artificial Intelligence" and not reviewed by the practitioner. The decision records no admission and names no system. The Panel added that busy clinicians may understandably rely on AI, but that opinions and references must be reviewed and verified before anyone relies on them, and it quoted the practitioner's own regulatory body to the same effect.

The Panel gave the medical opinion no weight. It noted that a second opinion from the same practitioner, filed in the companion case, contained similar errors involving different journal articles.

The veteran still succeeded. The Osteoarthritis Right Shoulder decision grants pain and suffering compensation at five-fifths under section 45 of the Veterans Well-being Act, effective March 1, 2026. The Panel relied on the presumption of fitness at enrollment, a service-related shoulder injury recorded on March 3, 1964, the veteran's testimony that the pain began around 1988, the occupational task statements, and the board's Entitlement Eligibility Guidelines for osteoarthritis.

Where it stands

The Osteoarthritis Right Shoulder review is decided. The decision carries the usual notice of a right to appeal to an Appeal Panel; because the veteran received full entitlement, an appeal on this claim would be unexpected, and none was located as of November 16, 2026. The decision in the companion case, which the Panel says involved a second opinion with similar errors, was not located. No regulatory step concerning the practitioner was located, and the decision does not say that the Panel made a referral.

The obligation it engaged

The Osteoarthritis Right Shoulder review engages Clause 3 of the Standard, independent authority. Every citation in an expert opinion has to be confirmed against the source itself. Here the articles appear to exist, which is the instructive part: the errors were in the authors and journals attached to them. A reference that resolves to a real paper can still be wrong in ways a tribunal will find, and a tribunal that finds two such errors will discount everything else in the document.

Clause 2, non-delegation, is the second clause. The Panel's objection was not that a system was used. Its objection was that the opinion, including its central timing statement, had not been reviewed by the professional who signed it. The 31-year error is a factual statement about the patient's own history, the one thing the treating clinician was best placed to get right.

The outcome carries a practical lesson for any expert. The veteran won on the service records, and the unreliable opinion contributed nothing. An opinion that cannot survive a reference check is not a weaker form of support. It is no support, and in a closer case it leaves the party who commissioned it without the evidence the party paid for.

Sources

What could not be confirmed

The Panel inferred AI use from the errors; the decision records no admission by the practitioner and identifies no system. The practitioner is named in the decision's exhibit list and is not named here for that reason. The Conclusion paragraph of the decision refers to the right elbow while the rest of the decision concerns the right shoulder; this record follows the heading and the operative decision. The companion decision, the opinion itself, and any appeal or regulatory step were not located as of November 16, 2026. The decision was read as hosted by the database, not on CanLII.

About this entry

This entry was checked against the primary document listed in the sources. The entry for Osteoarthritis Right Shoulder Review, 2026 CanLII 116891 records conduct and what the tribunal did. It does not characterize a holding and it is not legal advice. Last checked November 16, 2026. More matters like this one are collected under the expert witness matters. If something here is out of date, write to [email protected].

Top