Dive Brief:

• A law firm that used ChatGPT-4 to support its request for $113,485 in fees for helping a mother win specialized education services for her child received a stiff denial from the judge, who took issue with the technology’s reliability and with the firm’s lack of transparency in how it applied the tool in its request.
• The firm’s invocation of the technology, Judge Paul Engelmayer of the U.S. Southern District of New York said in his Feb. 22 ruling, is “utterly and unusually unpersuasive.”
• The Cuddy Law Firm, a specialist in education matters for special needs students, said it used the tool as a “cross-check” to support the hourly billing rates it sought but the judge suggested it need not have bothered. “Barring a paradigm shift in the reliability of this tool, the Cuddy Law Firm is well advised to excise references to ChatGPT from future fee applications.”

Dive Insight:

Cuddy Law Firm helped the mother win a pair of cases in which she alleged the NYC Department of Education failed to provide her child with specialized programming to which the student was entitled. The department offered to pay $54,300 to reimburse the mother’s legal costs. She rejected the offer and submitted to the court a request for the $113,485 in fees based on what the law firm said it was owed after calculating its rates and hours worked. 

The judge rejected the firm’s rates and hours worked as excessive, based on the complexity of the work involved and the firm’s past billing practices. It also dismissed surveys and other resources the firm referenced as part of its support for the fees. 

In dismissing the firm’s use of the generative AI tool, the judge cited two high-profile legal hallucinations the technology has been involved in, the first a case from last year in which the tool cited as precedent half a dozen cases that didn’t exists and the second a case earlier this year in which a brief relied on a nonexistent authority created by ChatGPT. 

In both matters, counsel were reproved for using the tool without checking the accuracy of the work. “ChatGPT proved unable to distinguish between real and fictitious case quotations,” the judge said.

The law firm didn’t help itself by disclosing little about how it used the tool to validate its billing rates, the judge indicated. 

“The Cuddy Law Firm does not identify the inputs in which ChatGPT relied,” the judge said. “It does not reveal whether any of these were similarly imaginary. It does not reveal whether ChatGPT anywhere considered a very real and relevant data point: the uniform bloc of precedent … in which courts in this District and Circuit have rejected as excessive the billing rates the Cuddy Law Firm urges for its timekeepers.

“The court therefore rejects out of hand ChatGPT’s conclusions as to the appropriate billing rates here,” said the judge, calculating the appropriate amount at $53,050, not quite as much as the NYC Department of Education originally offered to pay.