How Lawyers Can Prepare for the Future of AI in Legal with Bert Kaminski, Director of Legal for Google Cloud

Bert Kaminski is a seasoned legal counsel in the technology industry, specializing in advanced and emerging fields such as cloud computing, internet of things, artificial intelligence, data science & machine learning.  Bert is a frequent public speaker on these topics, and has served as a senior advisor to all levels of business and executive management in leading technology companies, such as Google, GE Digital and Oracle.

Bert serves as a Board Advisor to SafePorter LLC, a technology company specializing in privacy-centric data analytics and tools for diversity and inclusion tracking. Bert leads the Technology Law committee of the New Jersey chapter of the Association of Corporate Counsel (ACC) and previously served on the chapter’s board of directors and as its President.

As a Director within Google’s legal organization, Bert leads a team that supports commercial enterprise transactions for key industry segments of Google Cloud, including financial services, healthcare and the public sector.

Prior to joining Google, Bert was General Counsel and a member of the executive leadership team of ServiceMax Inc., a California based cloud application provider, where he was responsible for all legal affairs of the company including privacy and technology related legal issues.

Earlier, Bert was Chief Commercial Counsel at GE Digital LLC, a General Electric company, where he served as GE Digital’s privacy leader and specialized in the Industrial Internet of Things, cloud computing, and software licensing.

Bert began his in-house legal career at Oracle. Bert established and led Oracle’s Cloud Legal team, and as Assistant General Counsel, served as the principal attorney globally for Oracle’s cloud computing and information technology managed services businesses.

Bert served as an appointed member of the Privacy Bar Advisory Board of the International Association of Privacy Professionals (IAPP), and is also an IAPP Certified Information Privacy Professional for European data protection (CIPP/E).

A graduate of Fordham University’s joint J.D./M.B.A. program, Bert earned a certificate from the Massachusetts Institute of Technology (MIT) on the topic of Artificial Intelligence and its implications for business strategy. Bert earned a bachelor’s degree with honors at New York University (NYU), in the field of economics.

Jennifer Simpson Carr: Welcome, Bert.

Bert Kaminski: Thank you, Jennifer. Thank you for the introduction and I appreciate you having me on the show.

Jennifer Simpson Carr: For our listeners, you and I met, many years ago through Lowenstein Sandler and ACC New Jersey’s Cyber Day Program, which I saw recently is now in its seventh year. I want to thank you for being at the forefront of that initiative. I know it’s been very well received and I’ve been fortunate enough to keep in touch with you since then, including running into you at some of the ALM Legal Week shows.

Artificial Intelligence (AI) continues to improve work processes and assist lawyers with performing tasks, such as research, analyzing data, all of which can be valuable time-savers. Yet, these impressive advances in AI, specifically developed for use in the legal field, have some lawyers worried about security in the profession. So I’m excited that you’re joining me today to discuss technology’s role in legal and how counsel can adapt to these changes.

Bert Kaminski: Great. Looking forward to speaking about it.

What is intelligent contract automation?

I’ll answer it in this way: companies have, and law firms, have been implementing software and technology for many years and that’s ongoing and that’s sort of the normal tech automation that everyone’s doing. What intelligent contract automation is, is a little bit different. This is the next generation. This is where it’s not just software and not just servers and networking, but these are higher value, much more analytical computer programs that are backed by artificial intelligence that render deeper insights and inferences and make predictions, which are going to be useful for lawyers.

I’ve taken a couple of statistics here.

• • I noticed that Thomson Reuters a couple of years ago made a prediction that the AI service provider legal market was about $12 billion in 2017.
  • They predict it would go up to $85 billion in 2027.
  • I’ve seen other statistics where the annual growth of adoption of these technologies is about 35% compounded annually.

There’s a very, very big uptake by companies and law departments. Also, by the way, Altman Weil, a couple of years ago had a survey of major law firms and found that 48% of them were implementing AI contract automation and solutions to take over certain functions formally done by legal staffing lawyers and non-lawyers alike.

Jennifer Simpson Carr: 48% is a higher percentage than I would have guessed of law firms implementing this technology. It’s great to hear that such a high percentage are adapting.

What are some examples of legal contract on automation?

We’ve seen eDiscovery is a big one. Well, I’ll put it more generally as contract management and both on the transactional and on the litigation side. Programs that can generate contracts, that can review contracts, that can run analytics on those contracts, everything from doing red lines to doing much more sophisticated semantic understanding of what the terms and conditions are within documents. There are some capabilities to do on editing, extracting of information from natural language text, and contextualizing that so that attorneys can understand what’s in all of the documents.

As we all know, there’s been an explosion in the amount of documents and data in recent years, and there’s no way that any human can really do that. What a lot of these tools do is that they are able to review these documents, briefs, and contracts in a much more predictable, scalable, and fast way, better than any human can do. In sort of contract management, this is how lawyers obtain a better insight of what the commitments are from the extraction of this information, what the obligations are; renders better compliance, renders a better understanding of risk, allows for deeper and better negotiations in a way that does not require of full teams of associates to be doing.

Jennifer Simpson Carr: You’ve mentioned that this intelligence handles high-value, sophisticated types of work.

As we think about the adoption of this process, what are some ethical considerations as counsel are using these products?

If we would think of the code of ethics for lawyers, the American Bar Association (ABA) proposed Model Rules of Professional Conduct, and there’s the duty of competence, the duty of oversight, the duty of confidentiality, transparency to let your clients know what sort of tools you use. So, there are some ethical obligations. I do want to say is that some of these ethical, I will say considerations as opposed to risks because these tools are powerful, and they actually can enhance the legal practice. But the speed and scale of these tools do result in some potential overuse, perhaps by lawyers. A lawyer should not replace their own professional judgment by the output of a computer program, even one that is very sophisticated based on artificial intelligence and machine learning.

And, there is an automation bias. It’s a cognitive bias that we all have that if a machine provides us with an output, we tend to go with what it sets. We’ve all had recommendation engines and your news feeds and what to purchase and all of this, and there is a sort of an inclination to go with what it says. As machines and the AI-based tools are rendering more and more output for lawyers, better predictive analytics, more recommendations, more actions to be taken, more sorting of legal research, which is another key area, predictions on how juries will work, and judges will come out on different opinions; it behooves lawyers to really look very carefully at these tools and how they’re being used.

One needs to always have a ‘lawyer in the loop’, or more generally in the AI space we call it ‘a human in the loop’. When you have predictions or recommendations or potential actions to be taken by these automated tools, that there should be a part where the human lawyer steps in and makes the contextual decision on which way to go with it. So again, when you talk about the ethical parts it’s… If you look at Rule 1.1, the ABA Code of Conduct, that is the duty of Competence and included within the duty of competence is the duty to competently use emerging technologies in the practice of law. In fact, if one would not use these emerging technologies, that probably wouldn’t be a malpractice. A one has to use these in very thoughtful ways and also in compliance with the other ethical rules which I mentioned: supervise the use of the tools maintain confidentiality of client information, ensure its accuracy, and the like.

Jennifer Simpson Carr: You’ve already touched on my next question, which was some key considerations as these tools are deployed. Thank you for that. People talk about artificial intelligence and machine learning becoming too intelligent.

What’s the difference between narrow AI and general AI?

As of today, all of these contract automation tools that we see are focused on solving very specific problems, and we’ve seen examples. I mean, I play a computer chess game and computer chess has learned through a lot of machine learning and those games are really good. But that’s all they do. A contract automation tool, or a chess game, a chess-playing tool, or the AI that you’ve got in your iPhone for photographic optimization and all of this, they’re designed to do their one thing. They’re specified for the one problem that they can solve. And sometimes they do that well, by the way, sometimes you’ve managed it a little bit. But a general AI is akin to human intelligence, where AI automation tool can actually start adapting what it learns in one function and applying it to a completely different function elsewhere.

That’s not really at the point we are. There are some futurists who predict that with the rapid improvement of AI and machine learning capabilities, a more generalized intelligence can be had in 25 years even. Some will say that it never can be. And, there’s some that say this is impossible. But this is a little bit of the ways off and meanwhile, attorneys and users of these tools need to be a bit skeptical: that it may look very sleek and very fancy and very compelling what it gives outputs, but there may be some challenges in that, and one has to be aware of its use.

As we think about how these tools are being deployed in the various uses in the legal profession, do you think that human lawyers will be replaced by AI and contract automation?

It’s a tough question. According to the Altman Weil survey I mentioned before, that actually is beginning to occur. There’s a famous example, by the way, in some recent year: JP Morgan Chase implemented a contract review automation tool. And, in one recent year, they estimate it saved 360,000 billable hours of contract review work. That occurred within, let’s just say, seconds.

It may be that you don’t have a robot lawyer sitting next to you, but what happens is that as companies and law firms use these tools to automate the repetitive processes, the drudge work so to speak, the boring parts, the review of stacks and stacks of documents, trying to find details and make correlations and drawing out inferences, machines are really good at that. They will do that very quickly with less error and at a greater scale.

There will be some change. McKinsey & Company predicted that 23% of an average lawyer’s job can be automatable. Employment in the legal profession will start moving from associates and lawyers and legal staff, non-lawyers, doing contract review and contract management and process optimization: that type of thing can be automated. That may result in the reduction in the need for that kind of legal service, as it becomes automated. It requires lawyers to then up-skill and cross-skill and develop higher value type services to their clients to preserve and to enhance their careers.

How should counsel adapt their role in light of increasingly powerful AI-based tools and what are some critical skills that lawyers and counsel can start to refine?

I would say is that fundamentally, as a lawyer, don’t do what the machines can do and you try to do it better. If all we try to do is do more work and faster work, and we do more throughput and we try to be more efficient, and we try to engage in our peak performance, that can be automated. In fact, by the way, there was one AI researcher basically said, and I’m sort of extracting from this, but “if you can give a lawyer a playbook and say, ‘These are the rules, this is what you have to look for’ and the lawyer can read it and then can actually pick all that up within a few weeks, that type of function is easily automatable.”

What you want to do as a lawyer is get into areas of really depersonalized legal service, where you are giving your clients the things that machines can’t do. Machines are not very good when there’s ambiguity, when the data is amorphous or changing when there are two right answers, where there’s not a sense. Machines can’t tell what the… They don’t have a moral sense of what is the right answer or the right action, or the wrong actions: no right or wrong for machines.

Lawyers can, in this counselor sense, get to this way of really being this legal adviser in the sense of what we should do, why we should do it, how we framing questions and giving that higher level of concierge service because that’s what clients want. Clients don’t just want legal, analytical output; they want that type of higher-level service and value-add by lawyers: advice on how to solve problems and not trust the research end of it, and not just the analytical output and prediction. They want lawyers to take that, contextualize it, give recommendations that are personalized to the lawyers and to the client’s needs.

What resources would you recommend for individuals within the legal profession who are interested in learning more about the technological advances in the legal industry?

Well, I think I even saw you at one of these conferences at the ALM Legal Week in New York, which is a multi-day conference where AI-based types of legal contract vendors are gathered. They give presentations and demonstrations. It is a great conference for lawyers. Lawyers go and you can get CLE credits. I’m not advertising ALM Legal Week in particular, but that’s one type of industry legal conferences or legal tech conferences that one can go to see what the latest and greatest is.

I did mention that the ABA’s Rules of Professional Conduct do behoove attorneys to explore what is the latest and really what it is to adopt technologies that will be enhancing legal practice. It is incumbent on lawyers to be aware of what beneficial tools are out there. These types of legal tech conferences are one can also take other types of classes online, the CLE sessions that do talk about ethical issues and technological issues, adopting technology into legal practice. There are a lot of resources out there. ACC may have some of that as well among others. And, the ABA, of course.

Jennifer Simpson Carr: Thank you. We will certainly link in the resources that you mentioned so that our listeners can easily access those resources.

Burt, I want to thank you for joining me today and thank you to our listeners for tuning in. I’ve enjoyed our conversation and I really appreciate the feedback on how counsel can adapt in light of the tools that are changing our world. I’m excited to continue to watch the advances and appreciate you talking to me today.

Bert Kaminski: As am I. Thank you for having me. I enjoyed it.

Learn More & Connect

Bert Kaminski

LinkedIn: https://www.linkedin.com/in/bertkaminski/

Google Cloud: https://cloud.google.com/

Jennifer Simpson Carr

Learn more about Jennifer Simpson Carr

LinkedIn: https://www.linkedin.com/in/jennifersimpsoncarr/

Twitter: @jsimpsoncarr

Listen to more episodes of On Record PR