California fines lawyer $10K for ChatGPT-fabricated citations, warns: verify or face sanctions

A California appeals court fined an attorney $10,000 for filing a brief riddled with ChatGPT-fabricated quotations and issued a public warning to verify all citations. The ruling comes as California courts and the State Bar move to set AI policies amid a surge of fake case citations nationwide. Experts expect more incidents and call for sanctions, training, and education to enforce ethical AI use in law.

Key Points

A California appeals court fined attorney Amir Mostafavi $10,000 after finding that most quotes in his brief were AI-fabricated, and it issued a public warning to verify all citations.
California’s judiciary and bar are rapidly developing AI policies, with courts required to ban or adopt generative AI policies by Dec. 15 and bar ethics updates under consideration.
Incidents of AI-generated fake legal citations are rising sharply, with trackers noting dozens in California and hundreds nationwide, and even some judges have been caught citing fakes.
Research shows significant hallucination rates in legal queries despite widespread plans among lawyers to use AI, and detection may become harder as models grow more capable.
Proposed responses include fines, suspensions, mandatory training, and education to instill ethical, verification-first use of AI in legal practice.

Sentiment

The Hacker News community broadly agrees with the article's premise that sanctioning lawyers for AI-fabricated citations is appropriate. However, the dominant sentiment is that the $10K fine is insufficient — many commenters want significantly harsher penalties including disbarment. There is strong consensus that professional responsibility means lawyers must verify all submissions regardless of their source, with multiple practicing lawyers reinforcing this view. Some dissent comes from those who argue the real issue is negligence rather than AI itself, and from techno-optimists who believe grounded AI tools will soon make hallucination a non-issue.

In Agreement

Lawyers have an absolute professional duty to verify everything they submit to court; this sanction is justified and appropriate
Citation verification (shepardizing) has been automated for decades using tools like Westlaw and LexisNexis; there is no excuse for submitting unverified citations regardless of their source
The lawyer's unrepentant attitude about inevitable 'victims' and 'damages' demonstrates unfitness to practice and makes the sanction even more warranted
AI hallucinations in legal filings are a genuine, documented problem that courts need to address with sanctions and clear policy

Opposed

The $10K fine is far too low to be an effective deterrent; many commenters argue for disbarment, license suspension, or much higher penalties
AI tools with grounding capabilities like Lexis+ already prevent hallucinations, making this a transitional problem rather than an inherent AI limitation
The fundamental issue is professional negligence, not AI specifically — the same outcome could result from a sloppy human paralegal's unchecked work
HN users demanding harsh punishment for lawyers using AI while eagerly adopting it in their own coding workflows is hypocritical
AI could democratize legal access for people who currently cannot afford representation, though this hasn't materialized yet