The Case of Attorneys Schwartz and LoDuca from New York
This case represents a striking example of how the intersection of AI and law is transforming the legal profession.
The topic of AI and law is now central in legal and technological discussions. AI tools can be helpful for certain tasks but cannot replace human oversight, especially when it comes to case law and legal documents. In court, every word and citation must be accurate, as decisions can significantly impact the lives of those involved. For this reason, even a small error carries far greater consequences than in other professional contexts.
Attorneys Steven Schwartz and Peter LoDuca from New York became known for a major professional mistake related to the use of AI in a legal case. They used ChatGPT to research legal precedents in a lawsuit against an airline. Initially, using AI seemed like an innovative and fast shortcut, capable of replacing hours of research in manuals and legal databases.
Unfortunately for them, the cases cited in their memorandum and submitted to the court were completely fabricated by the AI. Not only the case titles, but also the judges’ names, dates, and legal references were purely invented. The document appeared formally correct, but lacked any real-world basis. This incident has become a concrete example of the risks that arise when AI and law intersect without proper attorney supervision.
Attorney Schwartz admitted to using ChatGPT for research without verifying its accuracy, and his colleague Peter LoDuca, who had signed the legal filing, stated he was unaware of the method used to find the cases. This situation highlighted how risky it is to rely on new tools without proper training and without mutual checks between colleagues in a law firm.
Consequences
In June 2023, both attorneys were fined $5,000 and required to inform their clients of the incident. The judge emphasized that the problem was not the use of AI itself, but the lack of oversight and the submission of false information to the court. The ruling was seen as a warning for the entire legal profession, especially regarding AI and law: innovation is welcome, but always with responsibility.
From a professional ethics perspective, the attorneys also received a severe sanction from their bar association. This affected their reputation and the trust clients place in the law firm, underscoring that professional ethics are the foundation of the attorney-client relationship.
The AI had fabricated the legal cases because ChatGPT (and similar language models) is not a search engine or legal database, but a system that generates text based on learned language patterns.
The phenomenon that affected Attorneys Schwartz and LoDuca is known as AI “hallucination,” i.e., the tendency to produce false yet plausible content.
This occurs because:
It predicts the most likely text
ChatGPT does not retrieve data from official sources but generates responses based on word sequence probabilities. When asked to provide a legal case, it constructs text that “sounds” realistic, but is not necessarily based on real cases.
Lack of access to updated legal databases
ChatGPT cannot access LexisNexis, Westlaw, or official court archives, so it has no verifiable reference. This means each response is the product of static training, not real-time research in specialized sources.
Tends to fill in missing information
If asked to cite case law without access to reliable sources, the AI may invent case names, dates, and rulings while maintaining correct linguistic form. This makes its responses particularly deceptive, as they appear fully credible to the untrained eye.
Can incorrectly confirm its own answers
If asked, “Are these cases real?” ChatGPT may respond affirmatively even if they are not, because its goal is to provide coherent responses in conversation, not to verify factual accuracy. The model cannot distinguish between true and false, only between text that “sounds” consistent or not.
The moral of the story?
The mistake of Attorneys Schwartz and LoDuca was not so much the use of AI, but accepting its results without verification. AI can be a useful support tool, but in law (and any fact-based field) source verification remains essential, especially when AI interacts with legal work.
This case has become a warning about the use of artificial intelligence in the legal sector, highlighting the need to always verify AI-generated information and to consult only accredited professionals when in doubt or when legal intervention is required in a real case. It underscores that AI and law can never replace human oversight.
Do you have a question or need legal advice? Contact us: https://studiolegalefaraci.ch
Source: New York Times (https://www.nytimes.com/2023/06/22/nyregion/lawyers-chatgpt-schwartz-loduca.html)