As I recently wrote, I attended the Summit AI New York last week (which, as the name suggests, was in New York City). Summit AI is a global conference and exhibition focused on the application of AI in business in general as opposed to legal in specific. The numbers from the 2-day Conference are impressive: more than 4000 attendees, 400 speakers across 11 different areas, and over 100 sponsors and exhibitors.
The Conference has been held for several years. Summit AI is a series of conferences that, in addition to New York, are held in London, Las Vegas, Singapore, and other locations. (There was one law firm with a booth by the way. Foley and Lardner, a full-service lawyer with some 26 offices, had a well-staffed booth and told me they are regular attendees).
I like attending nonlegal tech conferences because they often yield insights we don’t get in our legal tech cocoon. As one of my clients used to say, those of us in legal tech have to be careful lest we end up in a closet talking to ourselves too much. Indeed, as I wrote while at the Conference, I did get some new perspectives on how Gen AI can help tackle A2J problems.
So, after some time to reflect, what did I think of the Summit? There were certainly a lot of well-qualified speakers from well-known businesses.
The Conference tried to go too broad at the expense of not going deep
But the Conference tried to go too broad at the expense of not going deep. Different businesses have different perspectives on the use of AI and different ways of using it. These perspectives and uses are unique to each type of business. In trying to make sessions applicable to all kinds of businesses, some of the relevance and usefulness of the presentations are lost. Part of the problem was organization; instead of organizing the speakers by a business or specific issues, Summit AI was organized by “stages.” The stages were groups of speakers devoted to such things as Next Generation, Practitioners, GenAI, etc.
Another criticism: like so many AI based conferences, too many speakers talked too much about the how of AI and not enough about what uses to which the AI tools could be put. It’s nice to be told how to make a cake; it’s better actually to be given a piece to sample. Case in point: an OpenAI speaker gave a nice discussion about features and benefits. At the end of the presentation, the speaker asked the audience if they would rather me take questions or have a demo. DEMO. PLEASE.
I can’t really fault Summit AI. Every other Conference I’ve attended had the same problem. Also, I did not attend every session, so some speakers may have done a better job showing instead of talking.
Many of the AI related issues being faced by legal are also being faced by many other businesses
Despite the shortcomings, it still was a good Conference. I learned some things and confirmed that many of the AI related issues being faced by legal are also being faced by many other businesses. What were my top ten takeaways? Here they are in no particular order:
1. While there are concerns about the jobs that AI and GenAI will displace, it’s a fair point that many people in supporting roles in businesses and law firms can use AI tools to upgrade their skills and thereby provide things that businesses did not have access to before. This means, for example, that paralegals and law librarians can begin to do more things that the firm or client could not afford to have done before. Given the apparent shortage of legal talent in certain market segments (about which I have written before), this ability could be essential and replace some of the work lost due to automation and AI.
2. The need for workforce training about the risks and benefits of AI is paramount. As AI navigates to consumer devices, workers will become used to using AI tools and will inevitably want and will use these tools in the workplace. Businesses and law firms must accept this and ensure workers understand what the tools can and can’t do and how to protect the business and client. This need is especially true for lawyers and law firms. We have an ethical duty to appropriately supervise others who deal with client information.
3. For now, firms need to prepare their workforce before rolling out and implementing AI programs. If workers aren’t prepared, they will resist implementation, and the effort will be wasted. Relatedly, the starting point for any AI intuitive is to fully evaluate the tool. Otherwise, the tools won’t work and address pain points. This evaluatory need is critical since AI programs are quickly developing and changing. Businesses considering the implementation of AI programs need to start by evaluating the product and knowing what is offered because the field is developing quickly.
4. Humans must work with AI programs to get the best result. AI must be integrated with humans. This has been said so many times that it’s cliché. But what does this really mean? Where is this partnership going? No one wants to address these tough questions.
5. Thus far, AI is heavily used in dealing with supply chain issues, health care, and customer service. There was little discussion of its use in legal although one speaker pointed out that a pharmaceutical company’s legal department had “100%” adopted AI, whatever that means.
6. GenAI tools are really good for brainstorming but not for making a final decision about the best option and implementing that option. Humans still have to do that.
7. The data to which AI tools are applied is critical. You can’t just apply AI to any data and expect the tool to perform. It’s essential to first address whether the data has information that can yield an answer. Otherwise, the AI tool will not do what you think it will.
8. Why does the implementation of AI tools in a business fail?
- No clear goals and lack of success criteria
- Lack of expertise and understanding
- Poor data
- High cost of scaling
- Lack of trust
9. New AI tools are focusing on better voice and conversational interactions. The tools will get better at evaluating and providing information about photos and documents. The tools will be able to understand spatial relationships between objects, understand and react to audio nuances, and create video from text. The future is bright.
Everyone seems to agree that the future of AI is the creation of agents
10. Everyone seems to agree that the future of AI is the creation of agents. These agents will be able to do long-running and sequential tasks and multiple things at once. The agents can then act on the prompts and go out and do things. For example, you might ask the agent to create a demand letter, send it to me for review, send it to my adversary and diary it for response.
All in all, it was a good Conference, and I’m glad I went. If nothing else, it’s always nice to be in New York at Christmas time.