This week, the 10th annual AI Summit kicks off at the cavernous Javits Center in New York City. The promoters describe it as an event that “[b]rings together visionary leaders, innovators and technologists who are shaping the future of artificial intelligence.”

The Summit

Last year, there were over 5,000 attendees, over 100 exhibitors, and 350 speakers, according to the official website. This year promises those numbers will be exceeded. The keynote speakers include such notables as:

Matthew C. Fraser, Chief Technology Officer, The City of New York
Cecilia Kushner, Chief Strategy Officer, NYC Economic Development Corporation
Nitzan Mekel-Bobrov, Chief AI Officer, eBay
Anusha Dandapani, Chief, AI Hub, United Nations International Computing Centre (UNICC)
Jorge Reis-Filho, Chief AI and Data Scientist, Oncology R&D, AstraZeneca
Terry Doyle, Managing Partner, TELUS Global Ventures

The overall speaker lineup reflects AI’s extension from a tech curiosity 10 years ago to being embedded, for better or worse, across government, commerce, finance, and healthcare. Legal can’t afford to ignore this shift and its ramifications and risks.

I will be there covering for Above the Law and reporting what’s being talked about and perhaps what’s not.

But before we get to that, it’s worth pausing to see how far we’ve come in the decade since that first Summit in 2016.

The Past 10 Years

We’ve gone from AlphaGo beating a grand master at Go (2016) to LLMs (2022) to chatbots (2023) to agentic AI (2025) to speaking of AGI as a realistic possibility, all in one decade.

As for technology in general, here’s where we were in 2016:

 Amazon released the Amazon Echo Dot

 Slack was the hot new workplace tool with 4 million daily active users

 iPhone 7 was the newest phone (no Face ID, and first haptic home button)

 Windows 10 was still a relativity new operating system

 Netflix was starting to compete seriously with cable

The hottest play was Hamilton. Uber was becoming a real thing. In early 2016, a bitcoin was about $400 and WeWork was reportedly valued at $16 billion.

And in legal, in 2016:

 Clio published its first Legal Trends Report

 Most Biglaw firms were still debating whether to allow lawyers to use things like Dropbox for client files

 Document review was still almost entirely human-powered — predictive coding was “cutting edge”

 Legal project management software was considered “experimental” by most firms

We’ve come a long way.

Which makes this year’s Summit particularly interesting: has the AI industry matured along with the technology, or are we just experiencing the same hype cycles at higher and even dangerous volume?

What’s a Lawyer Doing at the AI Summit

I attended and wrote about this conference last year. First and foremost, it’s an opportunity to escape the legal tech conference bubble and see and hear what people in other businesses and professions are thinking about and doing. Like CES, which I’ll cover early next year, it often provides fresh perspectives.

Of course, with a conference this size, it’s sometimes hard to get a good handle on what’s really important. There are 11 tracks (the Summit calls them stages) with sessions that overlap. That makes planning challenging. Indeed, as I wrote last year, I have the feeling that the conference sometimes tries too hard and to do too much. That doesn’t make it a bad conference, just a challenging one.

Also, like most big conferences, the AI Summit is driven by vendors and exhibitors and is by design a bit of a rah-rah event celebrating AI. That also doesn’t necessarily make it all bad, but like CES, you have to take some of what’s said and exhibited with a grain of salt.

The Sessions

There’s a huge and daunting number of sessions. Many are highly technical and some are vendor specific. But many are educational focusing on where we are with AI and how AI platforms can be practically implemented.

Among other things, I’m approaching the Summit with three questions drawn from the recent series I co-authored with Melissa Rogozinski: Will the sessions acknowledge the crisis confronting the infrastructure required to support AI ambitions? Will anyone confront the verification paradox, the reality that verifying AI outputs often costs more than the efficiency gained? And third, will the conversation move beyond vendor enthusiasm to implementation reality? In that regard, there are several sessions that look particularly interesting that I plan to attend:

Who Owns Intelligence Wins: Escaping the AI “Rent Trap”
The AI Backbone: How Cutting-Edge Infrastructure is Powering the Next Wave of Innovation in Science and Industry
Betting It All on AI: C-Suite Confessions on Risk, Reward, and Reality
Venture Capitalist Matchmaking: Finding Your Perfect Fit
The Investor’s Crystal Ball: What’s Next?

I’m hoping that at least some of these sessions will get at the very things we talked about in our series: infrastructure requirements that may not be met as AI platforms expand, economic models that don’t quite add up, and the persistent gap between what vendors promise and what businesses can actually implement.

The question is whether presenters will confront these realities or perpetuate comfortable fictions.

What Else I May be Watching

I also plan to attend the City of New York’s presentation on using AI to improve access to justice. I wrote about what the City was doing last year and its impact; it will be interesting to see where the City is this year.

There’s also a smattering of sessions on AI and its impact on people and cybersecurity. There are sessions on determining ROI of AI, how it can be used to enhance storytelling and creativity, along with more philosophical sessions on the role of regulation. Of course, agentic AI is front and center.

I’ll also focus on sessions covering the tension between innovation and regulation, case studies in real-world deployment, cross-industry comparisons, training, and, perhaps most critically, sessions addressing hallucinations and the verification paradox we’ve been writing about.

Game Time

Yes, if it sounds like there’s more than I can possibly cover, you’re right. My conference plan isn’t set in stone. Like most of these conferences, I’ll make game-time decisions based on where the substance is and what I’m hearing.

It should be interesting. Will the Summit confront the questions about AI infrastructure, verification costs, and implementation reality? Or will it be another celebration of potential without accountability? Will sessions examine AI for societal good? Or will it be all AI for AI’s sake?

Stay tuned.

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.

The post AI Summit 2025: Hard Questions Or More Hype? appeared first on Above the Law.