208 | 💪 💸 Law firm invests $5M in AI and loses 6000 billable hours

In today’s Brainyacts:

1. Law Schools and AI

2. Harvey.ai shares a product overview

3. Presidential debate false info already in leading models and other AI model news

4. Robots get human flesh for their faces and more news you can use

   👋 to all subscribers!

To read previous editions, click here.

Lead Memo

👩‍💼 🏫 US Law Schools Reacting(?) to AI + Share Your Ideas on my Syllabus)

A recent survey conducted by the American Bar Association (ABA), involving 200 US law schools, sheds light on how these institutions are adapting to the integration of AI into legal education.

First the bad news - out of the 200 schools surveyed, only 29 responded! I mean come on law school deans. I am unsure if you were too embarrassed to disclose you are doing nothing on this front or were too busy designing AI-focused projects and curriculum (unlikely) to answer a survey - but we can do better.

So what did we learn?

• 16 stated they have “classes that are focused primarily on teaching students about artificial intelligence

• 27 stated that they are “[i]n light of the profession’s increasing use of AI tools, [they] are  . . . considering changes to [their] current curriculum.” The operative word here is “considering.” I hope they are considering what to do out of an abundance of ideas and options rather than considering whether to do anything at all.

Some did share ideas on what they are considering:

New Concentrations and Courses: Some law schools are considering or already have begun to add new concentrations, areas of study, or courses specifically focused on AI, its applications in law, and its ethical implications.

Incorporation into Existing Courses: Beyond new courses, there's a trend towards integrating AI tools and concepts throughout the curriculum, especially within skills courses, legal writing, legal research, and professional responsibility courses.

Encouraging AI Use in Experiential Learning: There's a move towards encouraging the use of AI in experiential classes, suggesting a recognition of the practical, hands-on benefits of understanding and using AI tools in legal practice settings.

Enhanced Technology Training: Several law schools mentioned plans to provide regular opportunities for students to receive training in technology, including AI.

Putting my money where my mouth is:

I previously shared that I am joining the faculty at Indiana University Mauer School of Law. This is a full-time teaching and building gig as I will be also designing and building a new executive education program for the school.

This fall I am offering the following class and it is oversubscribed (waiting list) by 100%!

Course Title: Practical Applications of Generative AI in Legal Practice (3 Credit Hours)

Course Description: This dynamic course is designed for 2nd and 3rd-year law students keen on understanding and applying generative AI (GenAI) in the legal domain. The curriculum is structured to impart practical knowledge and skills necessary for navigating the complexities of GenAI within legal contexts. Students will delve into the mechanics of Generative AI, exploring its operational framework, from foundational concepts of machine learning (ML) and deep learning to the intricacies of Large Language Models (LLMs) and their application in the legal industry. 

The course begins with an introduction to AI, highlighting the transition from basic AI to ML, and then to more complex systems like GenAI and LLMs. Students will learn the differences between unsupervised and supervised learning, and the nuances of discriminative and generative models, with a focus on how these technologies are shaping the future of legal practice. Lawyers must understand how the models are built, what they are built with, and how to control them.

A significant portion of the course will be dedicated to hands-on training with Generative AI models, emphasizing their adaptability through fine-tuning for legal-specific tasks. Students will gain practical experience in prompt design and learn how to leverage these models to maximize efficiency and accuracy in day-to-day work, legal analysis, and document generation.

The course will also provide a critical examination of the legal dilemmas posed by GenAI, including data privacy, copyright and intellectual property issues, the challenges of deepfakes as evidence, and the ethical considerations in AI deployment. Through case studies and real-world examples, students will engage with current and pressing legal debates, preparing them to address these issues competently in their future careers. 

By the end of the course, students will possess a robust understanding of GenAI’s working principles, be adept at utilizing them in various legal applications, and have a keen awareness of the legal and ethical challenges posed by AI technologies. They will leave the course not only with theoretical knowledge but with practical skills and a personal prompt library, enabling them to effectively integrate GenAI tools into their legal practice responsibly and innovatively.

🚨🚨YOUR TURN: What is one topic or idea that you believe I must include in this course? Reply directly to this email with your thoughts.

Spotlight

🎥 💄 Harvey.ai releases short product overview video

Many of you have asked to better understand Harvey. Recently they have been getting coverage for being in “stealth” mode, meaning they have not been press-friendly or shared much outside of the customers they are working with. I wrote about that here.

Well, they just posted a video overview and it will give you a glimpse of what you can use Harvey for. The video moves fast so I am sharing screen grabs after it.

Here are some screen grabs:

Here you see a robust prompt with several documents loaded.

Here, Harvey is responding the the above prompt.

This shows you that you can load numerous documents.

Another view of document loading.

Anyway, this is likely more than most have seen of the tool.

AI Model Notables

► OpenAI's ChatGPT and Microsoft's Copilot repeated a false claim about the presidential debate

► Amazon is secretly working on a ChatGPT killer

► Microsoft has tricked several leading gen-AI models into providing forbidden information using a jailbreak technique named Skeleton Key.

► OpenAI lands publisher Time in comprehensive content deal

► BUT! ChatGPT is hallucinating fake links to its news partners’ biggest investigations

► OpenAI develops AI “critics” that evaluate other AI model weaknesses

► Google partners with Thomson Reuters, Moody’s and more to give AI real-world data

► Google uses AI to add 110 new languages to Translate

► YouTube is trying to make AI music deals with major record labels

► Amazon investigating Perplexity AI after accusations it scrapes websites without consent

► Another news organization is suing OpenAI and Microsoft for alleged copyright infringement

News You Can Use:

➭ Human skin on robot faces: Japan researchers develop skin that robots can use

➭ Researchers just solved AI’s biggest conundrum showing LLMs running billions of parameters can operate on just 13 watts of power without a loss in performance

➭ AI a ‘fire accelerator’ for Big Tech antitrust abuses, Germany’s antitrust chief warns

➭ NBC to use AI version of announcer Al Michaels’ voice for Olympics recaps (yes, he is getting paid too)

➭ Israel to build supercomputer to keep pace in global AI race

➭ Morgan Stanley is launching an AI assistant called Debrief for its 15,000 wealth advisors, built using OpenAI's GPT-4 model. Debrief will attend advisors' client meetings via Zoom and take detailed notes, drafting summaries and follow-up emails to save advisors time.

➭ New AI scorecard can assess strength of state, federal legislation: The Electronic Privacy Information Center published a document that allows scoring of state and federal AI legislation. 

➭ Death, taxes, and AI: How generative AI will change accounting (with a key lesson for lawyers as to client service)

➭ Morehouse College will become the first educational institution to deploy AI teaching assistants throughout its classrooms starting in the fall.

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.