US Judge Disqualifies All Lawyers for Letting ChatGPT Fight Itself

The legal drama unfolded in a contract dispute between lawyer Tom Withers and the city of Aberdeen. Instead of doing actual legal work, the attorneys representing both sides decided to outsource their brains to generative bots.

The result was a glorious digital circus where lawyers cited completely fictional precedents, basically paying ChatGPT to argue with its own hallucinations. Legal observer Rob Freund aptly described the fiasco as a "comedy of AI errors."

U.S. District Judge Sharion Aycock quickly realized she wasn't reading actual law, but rather the creative writing of a confused language model. Fed up with wasting the court's time on imaginary court cases, she cancelled the proceedings entirely and kicked all four lawyers off the case.

Two of the attorneys received a two-year ban from practicing in the court. All four were slapped with fines ranging from $1,000 to $3,500 depending on how deeply they crawled into the AI rabbit hole. One of the disqualified lawyers genuinely apologized, admitting she let the AI write the entire document and had absolutely no idea what an "AI hallucination" even was.

This is far from an isolated incident, as courts across the country are dealing with a massive influx of fully generated lawsuits. Earlier this year, an Omaha attorney was suspended after submitting an appeal where 57 out of 63 cited legal precedents were completely fabricated by AI.

The dream of replacing highly-paid lawyers with cheap algorithms is currently colliding with the harsh reality that bots just make stuff up when they get bored. It turns out that delegating a multi-thousand-dollar lawsuit to a glorified autocomplete tool is a great way to get disbarred, fined, and publicly humiliated.

Source: CourtListener