Former Donald Trump attorney Michael Cohen used Google's Bard AI to generate bogus legal filings in a bid to end his court-mandated supervised release, his lawyer admitted in a federal court filing Thursday.
Cohen is currently on supervised release after serving half of a three-year sentence following his 2018 guilty plea for arranging hush money payments to porn star Stormy Daniels during Trump's 2016 presidential run.
Cohen's attorney Danya Perry wrote that her client thought Google Bard was a souped-up search engine and not a Chat GPT-like artificial intelligence tool when he used the service to generate three phony decisions to support his argument for the court terminating his post-release supervision, according to attorney Danya Perry.
"This is a simple story of a client making a well-intentioned but poorly-informed suggestion," Perry wrote in Thursday's letter to Judge Jesse M. Furman.
Cohen's lawyers filed the motion asking Furman to relieve their client of their court's supervision in November, but the eagle-eyed judge found notable inconsistencies in the three decisions cited as support for their argument.
One was an excerpt taken from a Fourth Circuit decision that had nothing to do with supervised release, while another quoted from a decision by the Board of Veterans' Appeals, an administrative tribunal that does not rule on criminal matters, the judge wrote.
The third case, according to Furman's order, "appears to correspond to nothing at all."
Perry said that Cohen believed the decisions were legitimate when he handed them to attorney David Schwartz, who she blamed for failing to vet Cohen's research before bundling it with the November filing.
Cohen's filing is reminiscent of the legal chicanery detected by federal Judge Kevin Castel in June, when he fined two New York lawyers $5,000 for using ChatGPT to generate false cases in support of his client's case against Colombian airline Avianca.
