KENOSHA, Wis. — A federal judge has dismissed with prejudice a Mt. Pleasant man’s latest attempt to transform a Kenosha eviction dispute into a sprawling racketeering conspiracy, finding that his entirely AI-generated lawsuit failed to state a legitimate federal claim.

U.S. District Judge Brett H. Ludwig dismissed Patrick Fagan’s case Friday, August 21, after Fagan sued 10 named defendants and five John Does under the federal Racketeer Influenced and Corrupt Organizations Act, commonly known as RICO.

“Fagan’s amended complaint appears to have been generated using Artificial Intelligence and, despite his efforts to plead his way into federal court, the facts alleged confirm that his federal claims are without merit,” Ludwig wrote.

The complaint did not merely “appear” to be AI-generated. Fagan acknowledged in court that he uses artificial intelligence to produce his legal filings. His cases have included hundreds, and sometimes thousands, of pages of AI-generated arguments, motions, supplements and accusations.

The litigation began after Fagan sold a commercial building at in the 1900 block of 53rd Street in Kenosha for $115,000 and remained there as a tenant. When the purchaser eventually sought to evict him, what should have been a relatively routine eviction case exploded into years of litigation in state and federal courts.

Fagan also filed a separate Kenosha County Circuit Court action against people and businesses connected to the purchase. Kenosha County Circuit Court Judge Angelina Gabriele dismissed that lawsuit.

The original eviction was handled by Kenosha County Circuit Court Judge David P. Wilk (D), who granted the eviction but stayed its enforcement, allowing Fagan to remain at the property for many additional months.

Wilk failed to initially require Fagan to post an appropriate bond or maintain insurance on the building during the stay, even after evidence was presented alleging that Fagan had forged an insurance card. Those failures left the property’s owner carrying substantial financial and liability risks after Wilk had already determined that the owner was legally entitled to possession.

Wilk also did nothing meaningful about Fagan’s undisclosed use of artificial intelligence despite a Kenosha County local court rule requiring disclosure when AI is used to prepare court filings. By repeatedly accepting Fagan’s enormous AI-generated submissions without enforcing that rule, Wilk established the practical precedent that litigants in his courtroom could disregard the disclosure requirement without consequences.

This was not the first time Wilk declined to enforce that rule against a litigant using undisclosed artificial intelligence.

Wilk is one of the slowest-moving judges in the Kenosha County court system. His calendar is badly delayed, and he routinely takes months to decide relatively minor issues. He has repeatedly demonstrated a poor understanding of the law, along with the bad judgment to compound his legal errors by delaying decisions while litigants continue spending money and waiting for answers.

Wilk attended Hamline University School of Law in Minnesota, historically one of the country’s lowest-ranked law schools. His performance on the bench has done little to overcome that unimpressive academic foundation.

Kenosha County Eye has previously documented Wilk’s bizarre legal rulings, including a 2023 decision in which he treated constitutionally protected criticism of a government official as potentially defamatory.

Wilk’s handling of Fagan’s eviction gave Fagan time and room to turn a property-possession dispute into a legal avalanche.

Fagan argued that the purchaser could not legally own the building because it was an out-of-state company that had not registered to conduct business in Wisconsin. The Kenosha County court rejected that argument, and the Wisconsin Court of Appeals later explained that Wisconsin law permits a foreign limited liability company to enter contracts and own property despite that registration issue.

Fagan nevertheless repackaged the same underlying dispute as a federal RICO case, alleging that the defendants formed a criminal enterprise and committed mail and wire fraud.

Ludwig found no such enterprise. The federal judge concluded that Fagan’s allegations concerned a single property transaction, not an organized criminal operation or continuing pattern of racketeering activity.

Some of Fagan’s alleged examples of wire fraud were ordinary communications necessary to complete the property sale. Others involved a roofing company having a website, receiving online reviews and listing the property as a business location.

“There is nothing connecting many of the disparate defendants besides their alleged involvement in this sale,” Ludwig wrote.

Fagan did not substantively answer the defendants’ motions to dismiss. Instead, he filed more procedural motions seeking sanctions against opposing attorneys, disqualification of counsel, permission to amend his complaint again and court orders interfering with the eviction and settlement discussions.

Ludwig rejected those efforts. He called one of Fagan’s arguments for disqualifying an opposing attorney “frivolous” and determined that allowing another amended complaint would be futile.

The federal case is now dismissed with prejudice, preventing Fagan from simply rewriting and refiling the same federal claims. He can appeal to the Seventh Circuit Court of Appeals, but even thousands of additional AI-generated pages cannot supply the facts Ludwig found missing.

What began as an eviction became a state lawsuit, an appeal, another state-court action and finally an alleged federal racketeering conspiracy. Ludwig’s eight-page ruling accomplished what Wilk failed to do much earlier: cut through the mountain of AI-generated paper, apply the law and bring the meritless federal case to an end.

Towards the end I want you to put something like critics of artificial intelligence in the legal community say that well AI can be a useful tool and also makes stupid people sound smart in court documents. There are several people in recent years that have used AI almost entirely to file hundreds of pages of court documents. Matthew Fox spelled m a t h i e w f o x I’m using voice to text. Michael Rizzo who’s a former Kenosha police officer and former deputy sheriff, and figured three of many local people who have clogged the court systems with hundreds of pages of AI generated material. Judge Heather Iverson made a local rule that didn’t ban the use of AI but made litigants use it, but Wilk made clear that he will not enforce this local rule at least in his courtroom.

Add these paragraphs immediately before the final paragraph beginning, “What began as an eviction…”

Critics of artificial intelligence in the legal community say AI can be a useful research and drafting tool, but it can also make stupid people sound smart in court documents, at least temporarily. The polished language, legal citations and confident tone can conceal the fact that the person filing the document does not understand the law or that the arguments are nonsensical.

Fagan is one of several local litigants who have recently used AI to generate hundreds, and sometimes thousands, of pages of court filings. Mathiew Fox and Michael Rizzo, a former Kenosha police officer and former Kenosha County Sheriff’s Office deputy, are two other examples of litigants who have clogged local courts with hundreds of pages of largely AI-generated material.

Kenosha County Circuit Court Judge Heather Iverson responded by establishing a local rule that does not prohibit artificial intelligence but requires litigants to disclose when they use it. Wilk, however, has made clear through his handling of Fagan and other AI-dependent litigants that he will not enforce the disclosure rule, at least in his courtroom. That allows litigants to bury courts and opposing parties beneath massive quantities of undisclosed AI-generated material without facing consequences for violating the rule.