Last year, recognising the rapid growth and countless applications of AI-generated writing tools, we reached out to in-house lawyers in our community to learn how they were integrating programmes like ChatGPT into their work.

We had an insightful conversation with Nick Tidmarsh, Legal Counsel at Third Bridge Group, where he shared his perspectives, use cases for ChatGPT, and advice for those interested in leveraging AI tools.

A year later, we reconnected with Nick to see if generative AI remains a valuable tool for him, how his use of AI has evolved, and his advice for those still contemplating the adoption of AI in their professional tasks.

Written by Nick Tidmarsh, Legal Counsel at Third Bridge Group

How has your use of AI evolved over the past year? Have there been any significant changes or advancements in the way you utilise AI tools?

The AI tool I use most in my own day-to-day work has remained overwhelmingly ChatGPT, as its uses are most relevant to my role.  While I have been impressed with other tools such as the new DALL-E model, their applicability to my role has been minimal, or their benefits have not been worth their cost. 

The ways in which I have used ChatGPT remain largely the same: as a stand-by assistant with a wide (but relatively shallow) mix of talents whose suggestions require review. From the IT side, I have found in the past year that ChatGPT particularly shines as a coding assistant which allows me to take full advantage of other IT tools I use in my legal practice, particularly Microsoft Office. 

From the legal research perspective, I still find AI less useful as a researcher than more traditional research tools like Practical Law, mostly due to their tendency towards light detail and hallucinations. But I think they still have a utility as a starting point to bounce ideas off and suggest routes of research (since they can process plain English summaries of issues and scenarios), until they become more powerful and focused.

What new AI technologies or features have you adopted in the past year?

The main new technologies and features are the ones that my company has been adopting as part of its client offerings. I work in the expert network sector, where companies often have substantial libraries of unique intellectual property. My own company has a large collection of transcripts of expert interviews, so there are obvious opportunities to leverage this distinct IP in a large language model.

So increasingly over the past year my team has been involved in the implementation of large language models in our products, such as an AI-enhanced search of our library. Plus, our clients are increasingly interested in feeding our IP directly into their own AI solutions, so AI-focused copyright licences are a new contract type my team has been building experience in this year.

Can you share specific examples or success stories where AI made a notable difference?

The most satisfying success has been ChatGPT assisting me in completely automating the initial drafting of my department’s most used contract templates. My coding knowledge is minimal, but it walked me step-by-step through upgrading our Microsoft Word templates so that the templates can be rapidly filled out, by clicking a few programmed buttons or with a simple user form. The time savings on these humdrum tasks have rapidly accumulated.

From past experience, implementing this would usually mean co-opting the resources of a very busy IT team, which can be difficult in-house. But I managed to easily implement the exact customisation and automation that I wanted completely solo, with the AI understanding my plain English requests, which accommodates my lack of technical knowledge. The result is that my small legal team’s drafting set-up is now as sophisticated as those of large law firms’ I’ve encountered.

What are your next steps or future plans for AI integration in your legal practice?

My company is trialling multiple resources for future integration into employee systems, taking an overall cautious approach due to info security concerns. So future integration into my workflows depends on what new tools are approved, and how my company’s compliance framework around AI evolves.

But AI is becoming more involved in my legal practice in unexpected ways. As AI becomes more integrated in my company’s and its clients’ business practices, an interesting result is far greater interest in copyright law amongst my non-legal colleagues. Copyright infringement and ownership was previously not a particular concern to some of them, as it seemed more of a matter for the Legal department. However now that LLMs can greatly facilitate copyright infringement by quickly processing a company’s valued intellectual property. Plus, copyright licences involving AI processing are a topic of increased interest to our client base, and discussions of intellectual property rights in the input and output have readily apparent commercial implications. So, the everyday relevance of copyright law is becoming more immediately obvious to my more commercial-focused colleagues through the prism of AI, and there is an increasing need for me to keep on top of legal commentary on the rapidly developing area of AI in the context of IP law.

Based on your experiences over the past year, what advice would you give to other in-house lawyers considering or currently using AI?

As the market continues to fill with a glut of companies latching onto the trend, I encourage stringent research on the actual benefits of an AI tool before subscribing to it. Not all tools are equal, and the better ones I have encountered integrated well-trained and focused AIs into pre-existing products which were already useful.

I would recommend earlier involvement in your company’s procurement process when AI suppliers are considered. This has benefits in being able to provide earlier input on compliance or legal risks from improper tools or usage, but also in increasing your own familiarity with the technology and your colleagues’ views on how to best leverage the tools, which are often easily transferred to your own work.

And linking back to my earlier answers, consider the low-effort and low-reward work taking up your time in in-house, and investigate how AI can automate it to free you up for more valuable and interesting work for your company.