Prakazrel “Pras” Michel, a founding member of the hip-hop group Fugees, is seeking a new trial after being convicted of fraud and foreign lobbying charges in April. He claims that his former lawyer used an experimental artificial intelligence program to write his closing argument, which resulted in a weak and ineffective defense.
The Trial and the Verdict
Michel was accused of funneling money from a Malaysian financier named Jho Low, who is wanted for his role in the looting of $4.5 billion from Malaysia’s state-owned investment fund. Michel allegedly used straw donors to make illegal campaign contributions to Barack Obama’s re-election campaign in 2012. He was also charged with trying to influence the Trump administration to drop the investigation into Low and extradite a Chinese dissident back to China.
The trial, which took place in D.C. federal court, featured testimony from celebrities like Leonardo DiCaprio and former Attorney General Jeff Sessions. Michel pleaded not guilty and denied any wrongdoing. He was represented by David Kenner, a prominent Los Angeles criminal defense attorney who has previously defended stars like Snoop Dogg, Suge Knight, and Tory Lanez.
However, after a two-week trial, the jury found Michel guilty on 10 counts, including conspiracy, witness tampering, and acting as an unregistered agent for China. He faces up to 20 years in prison for each count.
The AI Controversy
In a motion filed on Monday, Michel’s new lawyers argued that Kenner failed to provide effective counsel for his client. They claimed that Kenner outsourced trial preparations to inexperienced contract attorneys and did not object to damaging and inadmissible testimony during the trial.
But the most shocking allegation was that Kenner used an artificial intelligence program called EyeLevel.AI to generate his closing argument. According to the motion, Kenner had an undisclosed financial stake in the company that sells the AI tool and wanted to use Michel’s trial to promote it.
The motion said that the AI program produced a “deficient, unhelpful, and a missed opportunity” that prejudiced the defense. It cited several examples of how the AI-generated closing argument was flawed, such as:
- Starting with an apparent admission of guilt: “Ladies and gentleman, this case started back in 2012 when there was, as the government characterizes it, an effort to funnel money to President Obama’s reelection campaign.”
- Failing to highlight key weaknesses in the government’s case, such as the lack of direct evidence linking Michel to Low or China.
- Making frivolous arguments that contradicted the evidence or the law, such as suggesting that Michel was not acting as an agent for China because he did not receive any money from them.
- Using vague and generic language that did not engage the jury or persuade them of Michel’s innocence.
The motion also pointed out that EyeLevel.AI issued a press release on May 10, boasting that its “litigation assistance technology made history last week, becoming the first use of generative AI in a federal trial.” The press release quoted Kenner as saying that the AI tool was an “absolute game changer for complex litigation” and that it turned “hours or days of legal work into seconds.”
The Request for a New Trial
Michel’s new lawyers argued that Kenner’s use of AI violated Michel’s constitutional right to effective assistance of counsel. They said that Kenner was more focused on promoting his AI program than zealously defending Michel. They also said that the jury’s verdict was unreliable and based on a weak defense.
They asked the court to grant Michel a new trial or vacate his conviction. They said that Michel deserved a fair trial with competent counsel who would not use experimental AI tools to write closing arguments.
“Michel never had a chance,” they wrote. “The net effect was an unreliable verdict.”