228 |🧑🏻‍⚖️ ⚖️ AI used to write closing; Defendant seeks new trial

In today’s Brainyacts:

1. Using AI to draft closing argument

2. Elon went missing from AI100 - a website

3. Business/enterprise AI keeps growing and other AI model news

4. Nvidia: antitrust issue or not? and more news you can use

   👋 to all subscribers!

To read previous editions, click here.

Lead Memo

🧑🏻‍⚖️ ⚖️ AI Misstep in Court: Defendant Seeks New Trial Because AI Was Used to Craft Closing Argument

In the intriguing case of Fugees member Prakazrel Michel “Pras”, a novel and controversial legal issue arose when his defense team employed an artificial intelligence (AI) program to assist in crafting his closing argument, leading to an appeal for a new trial based on claims of ineffective assistance of counsel.

Michel was convicted on multiple serious charges, including conspiracy and acting as an unregistered foreign agent. In an effort to overturn his conviction, Michel’s argued that the AI-generated closing argument was deficient, citing several errors introduced by the AI, such as misattributed lyrics and a lack of engagement with crucial weaknesses in the prosecution’s case. The defense’s reliance on AI, ostensibly as a tool to ensure a thorough and compelling argument, paradoxically led to substantive errors and a perceived undermining of Michel’s defense.

The court’s decision to deny Michel’s request for a new trial rested on a detailed examination of the closing argument, both AI-generated and otherwise. Judge Colleen Kollar-Kotelly’s ruling (AI portion starts on page 40) was influenced heavily by the standard set in Strickland v. Washington, which requires demonstrating that counsel’s performance was deficient and that this deficiency prejudiced the defense.

Michel’s appeal highlighted the erroneous inclusion of lyrics mistakenly attributed to another artist, which his defense argued detracted from the credibility and effectiveness of the closing argument. However, the judge ruled that these errors did not meet the Strickland standard for prejudice—that is, there was not a reasonable probability that the outcome of the trial would have been different absent these errors.

This ruling brings to light several broader implications regarding the integration of AI in legal proceedings.

First, it underscores the potential pitfalls of relying on AI as a substitute for human judgment, especially in nuanced tasks like legal advocacy where contextual understanding and strategic judgment are paramount. The AI’s error in lyric attribution illustrates a fundamental risk: AI may lack the capability to fully understand and appropriately apply nuanced human knowledge and cultural context.

Second, the case serves as a cautionary tale about the importance of oversight when using AI tools in sensitive and high-stakes environments like the courtroom. The defense team’s oversight in catching the AI’s errors indicates a need for rigorous checks when using such technology.

Lastly, Michel’s case raises ethical and procedural questions about the extent to which AI should be involved in legal defenses. While AI can enhance legal practices by managing data and providing insights, its role in core advocacy tasks remains controversial and fraught with risks.

FWIW: The AI tool used was https://www.eyelevel.ai/law

Spotlight

🦉💪 I want you to use LLMs, not make myself look smart

In this edition, I’m offering a behind-the-scenes look at how AI can transform traditional content into dynamic, engaging experiences. As a writer, it’s easy to get stuck in the routine of producing dry, formulaic updates for your clients, especially when creating blogs, alerts, or emails. But what if there’s a more compelling way to present that content?

To demonstrate, I turned a short essay I wrote about Elon Musk's exclusion from Time magazine’s Top 100 AI influencers into a website—without a script, using tools like Anthropic’s Claude and their Artifact feature. This experiment shows how AI can quickly and efficiently reshape content in ways that capture attention and spark new ideas.

For those in client-facing industries, this approach might not be the answer to revamping your updates, but it’s a nudge to start thinking differently. It’s about breaking free from monotony and exploring new ways to present valuable insights that resonate with your audience. Check out my video for a peek into this process.

Key Bullets:

• Behind-the-scenes content creation: Watch me take a raw essay and turn it into a dynamic website using AI tools like Claude and Artifact—no script or planning involved.

• Relevance to your work: This is an example of how AI can inject life into content that’s often static or dull—particularly useful for those writing client updates, alerts, or blogs.

• Sparking creative ideas: Even if this isn't a direct blueprint, it offers inspiration on how to present your own work in more engaging, modern ways.

• Real-time process: In the video, I "work in public" for 10 minutes to demonstrate how quickly you can pivot from traditional writing to something more interactive and dynamic.

• Takeaway for readers: Use this as a prompt to rethink how you can bring your expertise to life in more interesting, engaging formats without overhauling your entire process.

• Content repurposing: By publishing both the video and the original essay, you’ll see how AI can help repurpose and reshape content across different mediums, giving it more reach and engagement potential.

🚨 BTW: here is the link to what Claude created: https://claude.site/artifacts/638aa108-a4d8-4b1e-90fc-d28f305d99df

AI Model Notables

► Microsoft plans to unveil the "next phase of Copilot" on Sept. 16.

► OpenAI hits 1 Million paid users for business (enterprise) versions of ChatGPT.

► Not to miss the enterprise party, Anthropic introduces Claude for Enterprise. Some interesting features but the one most legal users will be keen on is that you can collaborate with Claude with an expanded 500K context window—equivalent to hundreds of sales transcripts, dozens of 100+ page documents, or medium-sized codebases.

► YouTube is developing AI detection tools for music and faces, plus creator controls for AI training. 

► OpenAI could reportedly charge as much as $2K a month for high-end subscriptions.

► Ilya Sutskever, a co-founder of OpenAI, has raised $1 billion to launch Safe Superintelligence, a new AI startup focused on developing safe and powerful AI systems. The funding will be used to create a pure research organization with no intention of selling AI products or services in the near term.

► Facial recognition firm Clearview fined €30.5 million and banned from using ‘invasive’ AI in the Netherlands.

News You Can Use:

➭ These states are leading the way on enacting deepfake porn laws.

➭ UK regulator greenlights Microsoft’s Inflection acquihire, but also designates it a merger.

➭ Face the music': North Carolina man accused of $10 million AI-aided streaming fraud.

➭ US, UK and EU sign on to the Council of Europe’s high-level AI safety treaty.

➭ Did Nvidia receive antitrust subpoena from DOJ? Bloomberg reported it did. Nvidia says it did not.

➭ UK's first 'teacherless' AI classroom set to open in London.

➭ ‘Megalopolis’ trailer’s fake critic quotes were AI-generated, Lionsgate drops marketing consultant responsible for snafu.

➭ Indiana University researchers are leading a $7.5 million project to study how artificial intelligence (AI) can influence online communication, including misinformation and radicalization, and how these insights could help counter such effects.

Who is the author, Josh Kubicki?

Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.