We’re well into the roll out of Centre for Legal Innovation's Generative AI (GAI) Roundtables and this one on 27 March 2023, with legaltech developers and legal technologists, was our biggest yet!
It’s been fascinating to see how these Roundtables are evolving in real time in parallel with the blistering pace at which GAI is exploding! Generative AI (the company) is doing a wonderful job of providing real time info and examples of GAI and its application in many industries – do follow them on LinkedIn and elsewhere if you don’t already!
Huge thanks to Bianca Bowron-Cuthill for facilitating this Roundtable for us and to our wonderful contributors!
Fifteen big takeaways for me from this session:
1. It’s important to remember that generative AI is not just ChatGPT nor is it confined to text. GAI extends to images, video, audio, code, etc. – it is pervasive and because it is, it also provides opportunities to combine all these modes and reimagine how legal services/products can be delivered. For example, why not visually depict clauses/scenarios in a contract or build your own training modules? Should we continue to develop “algorithmic justice” as a means to settle e.g., small, commercial disputes and if so, how might we visually depict what outcomes look like? There’s a big focus on these multimodal capabilities in GPT4!
2. The conversations are now also starting to move from ChatGPT to all manner of working with large language models (LLMs).
3. As GAI unfolds, new skills, requiring lawyers and allied professionals to conceive work, undertake it and deliver it differently will need to be identified and people of all different experience levels will need to reskill or upskill. For those entering the legal world now, they already need different skills and legal education providers will need to be all over this and provide them.
4. Looking just a little way ahead, working with GAI will soon become such a fundamental part of legal work/all work, especially with its foreshadowed/expected wholesale integration into Office 365 (e.g. Microsoft 365 Copilot), if people cannot use it, they will be unemployable and/or will become redundant.
5. As ChatGPT (and similar competitor products) are integrated with firm/organisation data/information, that will be the real game changer. This integration will produce information (knowledge/knowledge hubs) that is accurate, verified and will also be customised for internal and external use. ChatGPT integration across all industries is on fire right now and quickly moving from being viewed as integration to the new App store!
6. ChatGPT provides a good starting point, something to build from. It has real and immediate application to tasks that are mundane and repetitive. It will force lawyers to rethink how and where they add value.
7. GAI provides a great catalyst for legaltech developers to revisit their product and service offerings and see how they need to change to continue to support their clients as that work also changes. Old products/approaches won’t work in a new GAI world!
8. GAI will result in time and cost savings for lawyers/clients but, the flip side of that is early career lawyers/allied professionals, will not be exposed to any or the same number of opportunities to repeat a task or observe how it is done. This could result in the lost foundational knowledge/understanding of what that task entails. Layers of these building blocks will be removed. Query how much repetition is needed to build a block? Are we focussed on building old blocks the old way when what we need is new ones built differently instead?
9. Query, if higher education (including legal) deploys GAI in teaching and learning, and removes some layers in the building blocks, will that translate to students in the future graduating less qualified, more qualified, or right qualified for a changing work environment?
10. ChatGPT can be a great tool for in-house lawyers when used in conjunction with a chatbot to create a low code/no code tool for internal clients. Working with ChatGPT will help identify the right questions to ask, will produce a first draft of responses (which can then be reviewed) to build a standard document for the organisation. It’s a new way of working for in-house counsel resulting in a faster turnaround time for internal clients and without (or with minimal) lawyer intervention.
11. GAI developers must continue to build in triggers and clear disclaimers/caveats for the generated responses i.e. more transparency about where the information is being drawn from and the consequent limitations on its accuracy. The inconsistencies and inaccuracies of what is being delivered right now (and with the understanding it is constantly improving) could mislead people who are unable to determine these limitations on their own. For example, GAI could be deployed to provide low/no-cost fast responses that could bridge the access to justice gap BUT, only if the information is accurate. Law students could also use it to assemble a list of cases on an issue but, ChatGPT has been found to “make up” or “hallucinate” cases. Unless these responses are correct, people could be misled by its plausibility.
12. Roles in the legal ecosystem will change and some will go. The role of the lawyer will focus increasingly on applying experience vs knowledge to a solution. Clients won’t need a summary of the knowledge/information, they will want legal experience applied to the solution in the context of the whole (holistic) issue/problem. A lawyer’s role will look more like a consultant i.e. bringing the right pieces together and interpreting what is built. Allied professionals will have the opportunity to add more value by collaborating on these solutions. These changes may make law as a career less interesting for some/many.
13. There are big issues to be resolved. For example, how will IP rights be claimed, on what, and how will that be assessed? How are issues of information/data governance and management being dealt with for public and private data used by GAI? How are ethical issues being addressed and by whom? Will each firm/organisation need to develop their own set of guidelines? Can the law keep pace?
14. Policies on the use/application of GAI are emerging in law firms but there are few of them. Key components so far are focussed on when GAI can and cannot be used (e.g. which data sets i.e. public vs private); if GAI is used in client matters, that must be disclosed to clients; when GAI is used, outputs must be manually reviewed and signed off before released anywhere.
15. It’s important to remain focussed on the human part of change and our fear of it. GAI has added new levels of expectation around speed of delivery of solutions (you can watch ChatGPT produce an output in real time), more to learn, more to manage…there is a human toll in all of this and it needs to be recognised and proactively managed.
Law school leaders will share their thoughts next. Given how many times legal education has been raised so far in this roundtable series, I can’t wait to hear how law schools/higher education are grappling with all of this.
About the Author
Terri Mottershead is the Executive Director of the Centre for Legal Innovation (Australia, New Zealand and Asia-Pacific) (CLI) at The College of Law. Terri collaborates internationally with leaders of legal businesses supporting them in identifying trends, developing strategies, and transforming their capabilities and practices to deliver legal services/products in the new legal ecosystem. She is the instigator, designer and developer in chief of CLI’s global initiatives, networks and programs including the Legalpreneurs Lab and its podcast series, The Legalpreneurs Sandbox. Prior to joining CLI, Terri was a practising lawyer, founded start-ups on three different continents, and established or led the in-house talent management departments for global firms and associations in Asia and the US including Lex Mundi, the Inter-Pacific Bar Association (IPBA) and DLA Piper LLP (US).