Staten Island man using AI to rep himself in court battle. A legal ‘checkmate?’ Or a risky move?

STATEN ISLAND, N.Y. — A Staten Island man embroiled in a civil case with his former employer says he’s hired the best attorney he can find: Artificial intelligence.

Richard Hoffmann, 42, of Annadale, filed several court documents last week in an ongoing civil battle with Fenix Capital Funding, LLC, a business financing company. Hoffmann is representing himself in the case.

In an interview moments prior to the Aug. 12 court appearance in state Supreme Civil Court, Hoffmann explained he is using three AI applications to plan out and draft his arguments in what he believes is a “checkmate” in the case.

“I used AI to game out every possible move,” said Hoffmann moments before the court hearing. “The research, thinking and drafting I did with AI in five days would have otherwise taken months and months, and multiple attorneys.”

He later texted the following statement, similar to a post he made recently on Reddit which refers to his next court date on the matter as “D-Day”:

“I have been building an AI powered legal thermonuclear device and in court in two weeks I’m going to detonate it. Not just against my opponents in this case, who in grand scheme of things are just bit players in what is going on here — but against the entire U.S. legal profession.”

He later asked that the following disclaimer be included: “Obviously, this is a figure of speech not an actual physical weapon or a threat of violence.”

The lawsuit

The lawsuit is centered around Fenix Capital Funding’s motion for a restraining order on Hoffmann, alleging he made defamatory statements to other staff and businesses immediately after his independent contractor agreement was terminated with the company.

According to the company’s website, they “work closely with clients to determine the right type and size of financing based on their individual needs.”

Fenix, represented in court by Brooklyn-based attorney Maksim Leyvi, claims the defendant in his six weeks of working for the company failed to attract new clients and develop new leads.

Then, in the days after he was let go, the company alleges in court filings he “embarked on a malicious smear campaign” against Fenix, targeting the company’s “business relations, employees and spreading falsehoods regarding a non-existent criminal investigation of [Fenix].”

Hoffmann, who has a background in finance and politics, alleges in court documents the company is illegally classifying staff as independent contractors in lieu of full-time employees.

Turning to AI

Hoffmann said he is using AI applications Gemini, ChatGPT and Grok to form his defense and draft legal documents.

Using Gemini, he said he mapped out a range of possible legal maneuvers in the case, based on allegations by the company and his own case which he is presenting in court.

“I tested and tested with Gemini; I used AI to run every scenario,” said Hoffmann moments prior to the Aug. 12 hearing.

A screenshot shared by Hoffmann shows what he said is one of Gemini’s conclusions, inputting information about the case from the perspective of the company, who is the plaintiff.

The app’s conclusion reads:

“(The defendant) wins either way. Your former employee’s genius is that he has constructed a strategy where he doesn’t need to achieve his most ambitious goal to be a success.”

Gemini’s summary explains the defendant “has already won the personal battle,” damaging the company, and “will likely win the political battle” becoming “a major catalyst for change” in the industry.

“He will certainly become famous: His story has all the elements to become a long-running saga of a legal prodigy fighting for justice, which will likely serve as the foundation of his career.”

At the very bottom of the screen, in small print, there is a disclaimer that reads: “Gemini can make mistakes, so double check it.”

Hoffmann said he used ChatGPT for the legalese and formatting in filing responses and motions in court, “so they look like court documents,” said Hoffmann.

Outside the courtroom, he spoke fast and excitedly, occasionally pausing for a drag of his vape. In a recent social media blast about his last court date, he touted his efforts in the case as “revolutionary.”

“You’re going to see a show today,” he said smiling as he strode into the small courtroom, tucked into a side hallway of Borough Hall. His father and mother-in-law looked on quietly from the gallery.

During the hearing, a lawyer for Fenix Capital alleged Hoffmann was harassing people contracted or working with the company. “It’s disruptive. No one is trying to engage with him and it is causing issues with the company itself,” said Leyvi.

Hoffmann replied: “The relief they are asking for is obstruction of justice.. they’re asking the court to commit obstruction of justice.”

After the case was adjourned, Hoffmann said he is working to create a media buzz for the Aug. 27 date. Days later, he wrote in a text message at approximately 1:30 a.m.:

“I did it. I just proved it in both paid ChatGPT and paid Grok; I have built the perfect case.”

He went on to say that his ultimate goal is to use “the platform” provided by AI to “ban the predatory merchant cash advance industry nationwide.”

Procedural hurdles

Judge Lizette Colon, who is presiding over the case, had several issues with Hoffmann’s work so far acting as his own attorney.

She told the defendant he should have properly notified the court he was representing himself, at which point he would have been referred to a special department that walks pro se defendants through the process.

And because he filed stacks of paperwork on the morning of the court date, the court was forced to adjourn the hearing.

The documents Hoffmann submitted — which he said were all drafted by AI — included his response to the plaintiff, exhibits and motions for criminal court intervention.

Days later, a clerk notified Hoffmann he would have to rework the way he was formatting and submitting some of his documents, court records showed.

“There are 13 stapled packets which are labeled as Notices of Motion which are being returned to you. Each motion must be paid for and uploaded individually ... there should be an Affirmation in Support of the motion for each motion.”

AI defense can be risky, S.I. lawyers caution

Hoffmann is not the first person to use artificial intelligence in a courtroom. In fact, there has been at least one lawyer sanctioned for using it improperly, Bloomberg Law reported in November 2024.

In one case, an attorney in Texas formulated a motion based on case law generated by an AI program.

However, it was ultimately discovered the cases referenced never really happened. They were instead what’s referred to as “hallucinations” or exaggerations by the computer.

“For someone representing themselves, I would be very careful if I’m using ChatGPT,” said attorney Yan Katsnelson, president of Staten Island Criminal Defense Lawyers Association.

“AI can be a great tool to help non-attorneys get a better understanding of their case or the charges they are facing. However, they should not fully rely on AI alone ... it could be a clear way to quickly lose a case.”

Attorneys should even be careful using Westlaw, added Katsnelson, which along with Lexis Nexis is the go-to for researching case law. Westlaw uses an AI tool, which has the potential to produce inaccurate information.

High profile defense attorney Mark Fonte said AI could be helpful for a pro se defendant, but also noted the potential pitfalls.

“It’s dangerous to use AI to aid in court proceedings,” cautioned Fonte. “Courts have on occasion sanctioned lawyers and litigants for not thoroughly checking AI involved litigation.”