Bold claim: AI is reshaping the practice of law, but not in the way many expect—this Oregon case shows AI can mislead, and the cost isn’t just accuracy, it’s accountability.
An Oregon attorney faced a steep consequence after relying on AI-generated material that proved to be false. The Oregon Court of Appeals ruled that Portland civil attorney Gabriel A. Watson presented two fabricated case citations and a bogus quote in briefs, prompting a $2,000 sanction from the state judicial department—$500 for each fake citation and $1,000 for the invented quote. The court described the situation as a “very grave” one, cautioning that while AI may seem like a time-saver for busy lawyers, it can paradoxically create a longer, more complex path to justice.
The decision marked a first for Oregon courts and underscored the risk of treating AI outputs as reliable legal authority. The errors came to light during a dispute in which Watson’s opponent, Charles Ringo, a former state legislator who filed the case himself, argued that Watson’s briefs relied on non-existent decisions. Ringo reviewed the documents, cross-checked citations, and confirmed that several authorities cited by Watson did not exist in the record. He even visited a Bend library to verify the sources and found the claims unsupported.
Watson attributed the mishaps to a “draft/placeholder” brief filed by his assistant and later apologized for the AI-driven mistakes. He urged the court not to sanction him, explaining the heavy workload of a solo practitioner and the temptation to lean on technology to advance justice for all clients. Nevertheless, the court rejected this defense, emphasizing that the legal system bears the cost of each erroneous brief and that AI, while designed to assist, does not substitute for careful, verifiable legal research.
Chief Judge Erin Lagesen highlighted that legal precedent remains the backbone of the judiciary, whereas AI operates on statistical likelihoods rather than verifiable truth. She rejected the notion that such AI misfires are merely hallucinations, clarifying that AI can generate nonexistent law according to its design rather than perceiving false information as a mental disorder.
The controversy extends beyond this case. Other Oregon matters involving AI-generated content have surfaced, including at least two federal cases where judges have grappled with sanctions and remedial actions related to AI-assisted filings. In one instance, U.S. District Judge Michael H. Simon declined sanctions after the party involved took corrective steps, signaling a cautious approach to AI-related missteps rather than outright punishment.
The broader takeaway is clear: AI can be a powerful tool in legal work, but it must be used with rigorous verification, transparent sourcing, and a robust understanding of its limitations. As technology evolves, the line between helpful automation and risky shortcuts becomes increasingly important to monitor.
What are your thoughts on AI’s role in legal practice? Is stricter oversight and mandatory verification the inevitable path, or should AI be embraced more fully with stronger safeguards? Share your perspective in the comments.
Zane Sparling is an award-winning journalist who has covered the Portland metro area for a decade. In 2025, he contributed to The Oregonian/OregonLive’s team that earned first place in a major journalism competition.
