Part of me wishes I could have used ChatGPT in law school. I hated reading judgements and writing case summaries.
Reading about Microsoft’s $10 billion USD investment into OpenAI’s ChatGPT got me thinking about how the legal profession could use it. Considering the majority of law firms operate in the Microsoft environment, it’s hard to see us able to run very far from ChatGPT.
This got me thinking about what we might see in the coming years, and what it might mean for law firms.
Thought 1: ChatGPT is efficient, but is it too efficient for law firms?
Thinking back on my law school days, I wondered how I could have avoided all the late nights reading and summarising cases.
So, I had a play in ChatGPT with everyone’s favourite ginger beer and snail story: Donoghue v Stevenson.
The original Donoghue v Stevenson judgement is about twenty five pages long, and I asked ChatGPT to prepare a case summary for me. It was able to prepare a summary that was just 3 paragraphs long.
The original judgement is about 21,939 words long. ChatGPT was able to summarise it into just 265 words. That’s a 98.81% reduction.
Assuming a fast reading speed of 160 words per minute, that takes the original judgement from a 137 minute read down to just 1.7 minutes. That’s a 98.76% time saving.
Efficient, right? For some, maybe too much.
Hypothetically, these reductions and time savings would mean that a Lawyer with a charge out rate of $300 an hour has just gone from recording over $600 in billable time, to about $30 (assuming you’re rounding it off into a six minute unit).
That’s a potential 95% reduction to the firm’s revenue for that piece of work, assuming it would actually get billed.
Thought 2: What will this mean for the business models of law firms?
I think the need to answer this question is dependent on whether or not the firm actually chooses to adopt the technology in the first place.
It’s obvious that firms want to either maintain or improve the state of revenue on their P&L. If the efficiencies I’ve described above can be achieved in reality, revenue would naturally take a hit, especially for those relying on the billable hour as their primary pricing model.
If a firm adopts this technology, I can’t see how the billable hour pricing model would be sufficient to preserve or improve their revenue given its reliance on time (and arguably, inefficiency) to work effectively.
Perhaps firms will counter this by adapting their pricing model towards more alternative fee arrangements such as value based pricing, if they haven’t already. Or, perhaps they’ll use ChatGPT as an opportunity to broaden their service offering, as a way to compensate for reduced revenue from the billable hour.
Thought 3: If law firms choose to adopt ChatGPT, how will they use it?
The potential of ChatGPT right now seems endless.
I actually asked ChatGPT to share how it could help lawyers do their job, and it listed contract review and analysis, legal research, document generation, knowledge management, predictive analysis and the automation of repetitive tasks as examples.
The above might eventually be true, but right now, the technology is still in its infancy. Until we see ChatGPT come out of beta testing, be trained with material that post-dates 2021, and we see some private entrants make their way into the market who can tick all the digital security boxes that law firms and their clients need, I don’t see the above happening too soon.
Just as well, it’s reasonable to expect legal regulators will eventually jump in and have their say. We saw this recently with governments in NSW and Queensland banning the use of ChatGPT in their public schools, and with regulators threatening jail time against DoNotPay’s CEO after an attempt to have the company’s “AI powered robot” argue a case in court.
In any event, after ChatGPT gave me the three paragraph summary of Donoghue v Stevenson, I asked it to explain the case to me at an eighth grade level. It did it quite well.
This immediately takes my mind to legal design. It makes me think of how ChatGPT can cut the fat around legal advice, legal documents, and how their impact is communicated to end users.
I wonder about the natural language processing capability of ChatGPT, how it can help serve end users from diverse cultural and linguistic backgrounds, and to what extent it can speak to them in a way that they understand.
I think this will play a significant role in enhancing access to justice, especially for those who encounter the justice system the most. Law firms aside, I think community justice centres and other NFPs in a similar boat have a lot to benefit from this technology.
Thought 4: How will those working in the legal profession compete with technologies like ChatGPT?
Although ChatGPT is still in its beta testing phase for the moment, the question of how those who work in the legal profession will be able to compete with the technology remains.
Surely I didn’t spend six years at university and $60K on a double degree with honours to have the processing power of my brain defeated by a chatbot in a matter of seconds.
Yikes.