Alabama Court Sanctions Lawyer for 12 AI-Made Cases

Alabama Supreme Court just sanctioned a lawyer for citing 12 nonexistent cases generated by AI.

AI Hallucinations in the Courtroom

The sanction follows a pattern that has emerged over the past three years, with attorneys repeatedly filing AI‑hallucinated citations despite warnings. A database maintained by Damien Charlotin, a senior research fellow at the Paris School of Advanced Business Studies (HEC Paris), lists more than 1,400 cases where courts have addressed AI errors. Charlotin’s AI‑powered reference checker, Pelaikan, promises to flag fabricated precedents before they hit the docket.

Trust Bias Amplifies the Problem

A February study showed that participants told AI guidance performed worse than those told human guidance, even when the advice was correct only half the time. The researchers found that people with positive attitudes toward AI were more likely to accept false guidance, a bias that does not appear with human advice. This effect underscores why even well‑meaning users keep trusting hallucinated outputs.

Cognitive Surrender and the Future

Wharton researchers found that giving users item‑by‑item feedback and cash rewards reduced, but did not eliminate, deference to faulty AI, a phenomenon they dubbed cognitive surrender. In a drone‑warfare simulation, participants reversed accurate classifications when a random‑feedback bot disagreed, illustrating how AI can sway life‑or‑death decisions.

If we ignore these signals, the legal profession risks sanctioning more attorneys, firing journalists, and, in worst‑case scenarios, making life‑or‑death decisions based on hallucinated data. Implementing systematic reference‑checking tools like Pelaikan is the next step; without them, courts will keep sanctioning attorneys and the profession will lose credibility.

Source: Why lawyers keep citing fake cases invented by AI
Domain: scientificamerican.com