International Legal Practice: What's ChatGPT Got to Do with It?

In Tomorrow’s Lawyers Richard Susskind predicts a world of online courts, AI-based global legal businesses, liberalized markets, commoditization and alternative sourcing, disruptive start-ups, legal training in virtual reality & a new range of law jobs. The book, released in its third edition last month, is essential reading for anyone interacting with the legal sector in coming decades (truly - highly recommend!).

Susskind’s outlook feels especially fresh at a time when Open AI’s ChatGPT is dominating most conversations in law schools and practice. Should AI generative software be used to support legal research and document drafting? Should clients expect their lawyers to be using this tech as part of diligent and thorough legal work? On the flip side, do lawyers have an obligation to disclose its use to clients? [Do I have an obligation to disclose to you whether or not AI generative software has helped me to draft this post?! (it hasn’t)] And most importantly for lawyers, what impact does this have on the billable hour metric?

Let’s take a deep dive into what ChatGPT is and what it can do, and then turn to what it should help lawyers do - especially international law practitioners.

What Exactly is ChatGPT?

ChatGPT is a form of generative AI built on top of OpenAI’s GPT-3 (Generative Pretrained Transformer 3) large language model. It uses statistical modeling, which allows the user to present open-ended questions and obtain responses in narrative format. While ChatGPT does not have access to the internet and only has access to information prior to 2021, its competitors, including Google’s Bard, do have access to the internet. Despite this important distinction, user studies show that ChatGPT is comparatively more advanced.

The user is able to regulate variability - the degree to which the bot generates the most or less likely next word in the sentence - by posing carefully crafted questions. This Q&A format makes ChatGPT especially attractive to all levels of users, and it also means that ChatGPT can respond effectively to complex inquiries. The better the user is at crafting questions, the better ChatGPT will be at providing the desired output from a stylistic perspective. Even more compelling, users can (and likely should) direct the length, tone, style, and ‘voice’ of the answers they hope to receive - again, the better the input the more useful the output.

Last week I received my first ChatGPT-generated email. It reflected regret and apologized for the sender’s failure to meet a deadline. I would not have known that the email was AI-generated, except it carried a ‘P.S.’ disclaimer that advised me the email had been fully written but ChatGPT, but it reflected everything the sender had wanted to tell me.

If I wasn’t already fascinated enough, I am now fully invested in unlocking the potential of this tech to support my work.

Progress Toward a Brave New World of Legal Practice

Why is it important to master and harness the potential of AI generative technology?

Depending on who you ask, ChatGPT is either leading the way towards revolutionary empowerment or threatening lawyers’ very existence. I discussed this topic with my students at GW Law recently and reactions ranged the gambit, from excitement about the prospect of being unburdened by busy work, to fear that AI will make junior lawyers obsolete. Several times, my students reminded me that ChatGPT-4 recently aced the bar exam.

This range of reactions shouldn’t be surprising. 高盛 predicts that generative AI could impact 300 million full-time jobs globally and automate 44% of legal jobs and 46% of admin jobs in the U.S. The upshot, according to the same study, is that:

worker displacement from automation has historically been offset by creation of new jobs, and the emergence of new occupations following technological innovations accounts for the vast majority of long-run employment growth. The combination of significant labor cost savings, new job creation, and higher productivity for non-displaced workers raises the possibility of a productivity boom that raises economic growth substantially, although the timing of such a boom is hard to predict.

But with the pervasive billable-hour and concerns over professional liability & ethics, can AI technology really automate 44% of legal jobs? Susskind reminds us that the legal profession has remained virtually unchanged for the past 100 years - tasks that lawyers do today are more or less the same as those lawyers did a century ago. It’s thus unsurprising that lawyers are often resistant to even basic reliance on digital tools including ones that automate administrative tasks.

At the very least, this state-of-play elucidates the value of niche boutique law firms like mine. The ‘lean law’ movement has been steadily gaining traction - prioritizing efficiency and value over merely the quantity of hours worked. The flexibility of this business model has also been harnessed by practitioners who advise on and litigate on the international plane. In this vein, alternative fee arrangements, which facilitate greater predictability for both clients and their legal providers, are likely more compatible with the future of the profession.

Okay, Great. So What can ChatGPT Help Me Do?

My personal expertise is definitely at a nascent stage. My experiments and the deluge of articles providing best practice and jail-break hacks confirm that successful implementation requires effective framing of the query and maintaining realistic expectations for the responsive output.

A few observations:

First, there’s definitely potential for AI-generative tools to increase nuance and accuracy in research. It can search through vast amounts of data and to provide a user with the most relevant results. This tech provides a means to obtain ‘quick answers’ to legal questions before diving into complex research - a method that is much more powerful than your run-of-the-mill keyword searching because it immediately directs a user to more accurate and relevant results without the upfront investment of a lot of time. The accuracy of the results stems from the tech’s ability to understand and process natural language, which is especially useful for complex legal queries. This methodology can be applied to a variety of tedious legal tasks such as researching case law, finding legal documents, or verifying facts. It may also facilitate greater access to legal materials that may only be available in other languages. Relatedly, AI-generative tools can guide and ensure consistency/coherence in document management and e-discovery.

If lawyers are able to reduce the amount of time they spends on mundane tasks by automating (in whole or part) document retrieval, production, summarization, and analysis, then those same lawyers - who are now basking in free time - can reinvest their time into supervising the automated processes and, among other things, ensure that the tech is implemented in a manner that is compliant with relevant ethical and professional guidelines.

Second, successfully implementing AI-generative tools into a legal practice frees up lawyer time for strategic thinking. In fact, with the support of generative technology lawyers can harness a treasure trove of data, including prior legal submissions and precedent case law, to gain insights on the likelihood of success of various arguments and lines of thinking, as informed by global data sets and different legal traditions.

To be clear, because AI-generative technology cannot ‘think’, it cannot necessarily help with the legal interpretive process. But if the tech is able to queue up all of the right bits of information for lawyers (including the ways in which treaties, contracts, etc. have been interpreted in the past), then lawyers can directly move to weighing the strengthens of different approaches and strategic decisionmaking. This is especially compelling as courts and tribunals around the world continue to engage in various transparency initiatives to make their proceedings and decisions more readily available to the public.

Third, the Q&A format of ChatGPT and Bard is refreshingly easily. Anyone (including lawyers) can quickly gain fresh insights on emerging topics, new developments, and changes to stay on top of industry news & practice developments with ease. This low threshold of effort truly means lawyers - especially those engaged in international practice where there is always too much to keep up with - will be better informed on big picture developments and, thereby, likely more effective.

But most news coverage on ChatGPT is focused on its content generation capabilities. Can lawyers use it to ghostwrite? Maybe. Its prompt-based system is certainly a boon for lean law because it can support the creation of first drafts and outlines. And recall that ChatGPT4 was able to pass (ace!) the bar exam.

I can see the tech being useful for preparing first drafts of contract clauses with very precise and clear prompts. But I don’t at the moment foresee ChatGPT spitting out winning legal briefs and submissions that are ready to be filed before courts and tribunals. However, if the tech is able to generate passable first drafts and skeletons this can help lawyers get on with the drafting process and avoid the dreaded time that is spent staring at a blank screen with a blinking cursor.

On the other hand, AI-generated text may be passable when the goal is not necessarily persuasion but simply generation of content volume (e.g., legal marketing, social media, etc.) - so again, another win for lean law!

Risks remain. The lack of controls and unknowns with respect to data privacy should give lawyers pause. Italy is a great example of a jurisdiction that has approached ChatGPT with measured caution. I will leave for another day musings on whether we are on the cusp of AI-powered decision-makers/judges.

What do you think? Tell me in the comments!

Post Script

• I am beyond overwhelmed to be sending the first substantive edition of this newsletter to over 2,000 subscribers across Substack and LinkedIn. I have readers on all of the continents (except Antarctica) and in dozens of countries. Thank you for your support, confidence & above all, friendship.
• Check out the Amazon preview of Richard Susskind ’s Tomorrow’s Lawyers (newly available in its third edition).
• If you like podcasts, check out the recent episode of Wolters Kluwer: International Arbitration & Mediation ’s International Law Talk podcast where I interviewed Claudia Salomon , President of the ICC Arbitration , on technology in arbitration & various disruptions and opportunities that are anticipated to shape the future of dispute resolution.

Like What You See?

To join the interactive community & gain full access to the newsletter subscribe on the?website.

Subscribe now

You’ll never miss an update, and I promise your inbox won’t be flooded. My contributions will always be timely & insightful. I am excited to exchange ideas & build a community on this new platform!