Please share your background and experience in practicing law with a focus on AI. How has your career evolved as this technology has advanced? My specialty is open source software licensing, an area where I have been building expertise for nearly 30 years. I started out as a general tech transactions lawyer, with an emphasis on software licensing. Because of that, I have always focused on the outer boundaries of copyright protection, such as with open data and hardware licensing. Also, I represented Google in Oracle America v. Google, which went all the way to the Supreme Court and took over 10 years to litigate. When that case started, issues centered on open source licensing, but by the end, they were about the limits of copyright protection for software–so I had to develop deep expertise in software copyright. While my practice these days is mostly focused on open source software licensing, the issues in AI are quite similar, and today, I am spending a lot of my time counseling on AI and ML, particularly generative AI. What is happening today in generative AI is very similar to what happened in the early 2000s with open source licensing. While the lawyers are figuring out whether to prohibit it, use of the technology has forged ahead, and can't be stopped. So the best lawyers can do today is develop best practices for use of AI.
What are some potential benefits and challenges that lawyers and law firms should be aware of when using these tools? Generative AI can be hugely valuable. It can write first drafts of memos or other communications, help with research, and get ideas for your work product. You can even ask it for correct Bluebook format! The benefits are immediately obvious once you start playing with these tools. Most of the chatbots are not good for final work product, but they are a great help at speeding up drafting.
The risks are still being analyzed. If you use AI via a cloud service, how will your confidential information be treated? If you ask it to write patent claims, is that a publication? If you register a copyright for work you produce using the help of AI, do you have to describe the nature of the AI authorship in your registration? The one I have focused on the most, so far, is whether the output of models is infringing on the copyrights for the inputs, as alleged in the January 2023 class action against Stability, DeviantArt, and Midjourney, Inc. I think that proper model training is fair use, but it may be years until we have clear legal guidance on that question.
In the meantime, I worry that rampant litigation will only benefit the lawyers. Big defendants have a war chest to fight complicated fair use cases, but smaller players can be destroyed by large legal bills, or by lawsuits scaring off investors. If that results in industry consolidation, we will all suffer from less innovation and a narrower constituency to set policy on use of AI.
How can AI tools be used by in-house legal teams ethically and effectively? What are some ways that AI has changed your legal practice? Very soon, we will be seeing new products that filter model inputs and outputs. These filters may be built into a tool, or provided separately. For the time being, in-house counsel should develop policies for AI use, but these policies must be updated frequently as the understanding of risks develop.
What excites you the most about the future of AI? Any general thoughts you would like to share? New technologies are always scary. But I think fears of AI replacing humans is hugely overblown by the media to get people's attention. Here is an example for the business lawyers. About 40-50 years ago, comparisons of versions of a contract were done by hand. Redlining was a process done by law clerks, who painstakingly read documents side by side and annotated the changes. As Wikipedia notes, “Included in this process were the potential for human error and the expansive administrative time necessitated by this arduous process.” So I want to ask the transactions lawyers, would you like to go back to the time when redlines were done by hand, so you will have more work to do?
I don't, either! Today we have word processing programs that automatically create redlines for us. They are extremely accurate and fast. And none of us wants to go back. But this technology did put some law clerks out of business, didn’t it? Not at all, law clerks just began to do different tasks, and probably less tedious ones.
This is how almost all revolutionary technology affects a workforce. It doesn't replace humans, it changes what we do, and mostly for the better. After all, we can't all be blacksmiths.