August 25, 2026
WHILE AI TOOLS CAN BE useful at times, recent cases highlight the dangers of using them for claims and appeals.
Here are five recent BC cases in which workers damaged their own claims by using AI.
Rolf Harrison, Workers’ Advocate
AI legal
The worker had an AI assistant draft his appeal submission and submitted it without any input from a lawyer or professional.
Only problem: the AI misrepresented law and policy, cited fake legal cases, and overall, made his case worse. In their decision rejecting this case, the appeal tribunal specifically stated that AI is not a reliable tool for legal research, warning that workers and employers who provide AI-generated submissions may be penalized with costs.
AI science
The worker asked an AI chatbot whether the nature of his employment caused the arthritis in his back. The AI chatbot said yes, citing scientific articles, so he sent it to WorkSafeBC, which was deciding his claim. The only hitch: the articles did not support his case. They supported a link that fell far short of what the law requires to accept an injury. WorkSafeBC used it against him, denying his claim.
AI evidence
The worker believed his doctor had misdiagnosed his work injuries and was fighting WorkSafeBC over the issue. He uploaded his claim records to an AI chatbot and asked it to create a first-person account of how the work accident had caused additional injuries that his doctor missed. The results seemed perfect, so much so that he wanted to submit them as his own personal statement.
One snag: it was make-believe. The AI made up symptoms and misrepresented medical findings to create the narrative he wanted. He nearly committed compensation fraud.
AI medical
The worker was on the fence about the surgery recommended at WorkSafeBC’s clinic. He uploaded his medical records to an AI chatbot and asked it to identify potential medical errors the surgeon had made. The AI identified “troubling conflicts” between his medical records and “missing medical investigations”, convincing the worker to decline surgery. Of course, it was false and misleading “AI medical slop”. But the worker’s mind was made up, to the detriment of his health and his claim.
AI rewriting
The worker used AI to rewrite an important email to WorkSafeBC. He wanted to make it clearer and more professional.
Only problem: the issue at hand centered around his credibility. Suddenly this email didn’t sound like him. It was polished. It didn’t have all the typos and exclamation points like in previous emails. It lacked authenticity, causing the adjudicator to put less weight on it.
Takeaways:
- Use medical opinions from real doctors, not AI versions.
- Never rely or submit on AI-generated medical advice. Discuss it with a doctor.
- AI is unreliable for legal advice and submissions. Discuss AI research with a legal advisor.
- Never submit new AI content as if it were your own evidence.
- AI-rewriting can reduce your believability. Your own words are worth more weight than the AI-refined version.
- Protect your privacy. Most AI tools don’t protect your privacy when you upload private medical or personal information.
By Rolf Harrison,
Workers’ Advocate