As AI continues to evolve at an unprecedented pace, should the legal profession consider how these changes could redefine its core functions and values?
The post The Nvidia CEO’s Keynote At CES: What It Means For Legal appeared first on Above the Law.
Nvidia CEO Jensen Huang at CES 2025. (Photo by Artur Widak/Anadolu via Getty Images)
I see trees of green, red roses too. I see them bloom for me and you. And I think to myself, what a wonderful world.
— Louis Armstrong
There is a school of thought that AI and its use has gone about as far as it can go for now. The theory is that future applications, especially in the workplace, will basically be minor iterations of what we have now.
In the legal world, this thinking morphs into the idea that AI won’t significantly impact what lawyers do because it can’t do the work that really matters to clients. Or AI is nowhere near matching the intuition and gut instincts of experienced lawyers and won’t any time.
Of course, not everyone agrees. Just yesterday, January 6, Sam Altman, the OpenAI CEO, stated in a blog post that OpenAI knows “how to build AGI (artificial general intelligence) as we have traditionally understood it.” Altman also predicts that AI agents capable of autonomously performing certain tasks may start to “materially change the output of companies” this year.
Jensen Huang Keynote
That same school of possibility thinking was expressed last night in the opening CES Keynote by Jensen Huang, CEO of Nvidia. Nvidia is the world’s largest semiconductor company and a dominant AI hardware and software supplier.
Huang is a popular speaker mainly because he has mastered the ability to talk to a massive audience just like he is speaking to you across the kitchen table. (All the public speaking instructions tell you to do that, but it’s easier said than done.) Huang did it for over an hour last night, captivating the audience even though he sometimes talked in technical terms that were over my head.
Huang traced the evolution of AI and Gen AI, which now understands, translates, and generates images and text. Huang explained how neural networks and machine learning are advancing. He showed an AI computer generated video that was completely realistic and indistinguishable from a real life video taken in real life with real people. Huang told us the AI video was made by inputting a limited number of pixels from which the AI program inferred and deduced the pixels that needed to be added to generate the finished content.
Huang explained how the amount of data available to AI programs will increase exponentially over the next few years. This increased data can then be used to better train the AI models, enabling them to do more and more. Humans can also reinforce this increased learning, leading to even more exponential AI growth. And, Huang theorized, the AI program itself would learn how to improve itself. “In the future [AI} is going to be thinking. It’s going to be internally reflecting, processing. … And it’s interacting; it’s taking the problem you gave it, breaking it down step by step.”
Huang believes we are just beginning to see what sophisticated AI can do. We are moving, he said, from Gen AI, where computers create content, to agentic AI, where AI agents can actually do things without being given detailed instructions. These agentic agents will become invaluable digital employees who will do things on our behalf with little prompting. They will become research assistants, create and act on sophisticated weather forecasts, analyze and make traffic decisions, and monitor manufacturing processes, for example. Huang did not say what the humans would be doing, by the way. In the legal world, AI agents could do such things as better automate legal research, undertake drafting, or even formulate litigation strategy.
The next step, according to Huang, will be physical AI, where AI understands the physical world and how objects interact. He gave an example of a ball rolling off the table, and opined that physical AI would understand that the ball didn’t disappear but simply fell to the ground. Developing this kind of AI requires massive video input and investment in training. But it will lead to considerable advances in robotics as machines understand the physical world much like we humans do now.
What Does All This Mean?
I’ve learned over the years to take much of what’s said at CES with a grain of salt. All too often, its wishful thinking designed to get attention more than reflect reality. But Huang and Nvidia have an imposing track record that cannot be taken lightly.
If they are right, where does it leave us? In particular, where does it leave the legal profession, which for years, has touted and relied upon human communication, persuasion, and interaction? Our business is believed to be a particularly human one, for better or worse.
A Wonderful World?
I worry what will happen as the distinction between what is real and what is computer created blurs. What happens when it really no longer matters whether something is real or not? What does that do to evidence? What does that do to fact-finding?
Are we prepared for a world where a computer can come up with a better solution than a human? Where an algorithm can reach a faster and better result than a human judge or jury? When an AI program can construct a more compelling argument than a real life lawyer?
What will happen if legal AI agents can do much of the work that keeps law firms’ personnel busy? What will lawyers be doing in five years? Ten?
As I sit at CES this year and listen to AI’s possibilities, I can’t help but think that legal is a bit like the proverbial ostrich sticking its head in the sand. We stew about AI. We try to demonstrate why it can’t and won’t work in legal. We try to convince ourselves that AI has gone as far as it can go so we don’t have to worry that our law cocoon might soon burst.
Maybe the naysayers are right. Perhaps change won’t happen. But if there is even the slightest chance Huang and the other AI evangelists are right, the changes AI brings and brings quickly could upend our profession from top to bottom. And other than folks like Cat Moon at Vanderbilt University, Andrew Perlman, dean of Suffolk University Law School and David Wilkins at Harvard Law, there seems to be precious few thinking about what those changes might mean.
To paraphrase Steve Jobs, AI has no respect for the status quo. You can quote it, disagree with it, glorify or vilify it. But the only thing you can’t do is ignore it, because AI will change things.
Stephen Embry is a lawyer, speaker, blogger and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.
