Since ChatGPT’s launch in November 2022, the ‘end of lawyers’ narrative has found new life. From Goldman Sachs predicting that 44% of tasks performed by corporate lawyers would be automated by Artificial Intelligence (“AI”) to GPT-4 passing the bar, AI has put the legal profession on notice. Even the infamous case of a lawyer who was caught submitting a brief to the court written by Chat-GPT citing hallucinated cases likely did more to raise awareness with lawyers than it did to pour water on the idea of AI automating legal work. After all, Chat-GPT was impressive enough to convince a practicing attorney to use it to do their job. Top VC funds are pouring money into AI-driven startups catering to law firms, with Harvey (AmLaw 200 firms) and EvenUp (personal injury firms) both newly minted as unicorns in recent months.
AI will continue to reshape both the practice and the business of law, and that’s a good thing. Over the last 20 years, despite an increase in the number of lawyers, the number of people who must navigate disputes without legal representation has not reduced. In 75% of civil cases in the US state courts, at least one party does not have legal representation. And 92% of Americans cannot afford legal representation for their civil justice problems. Most small and medium businesses are legally naked because they can’t afford to hire counsel and face significant risks as a result. The global economy is rooted in psychology and the rule of law is the bedrock of our prosperity. But, today, most people do not feel like the rule of law is working for them. As the Chief Justice of the Michigan State Supreme Court, the author was confronted with this gloomy reality every day.
It is encouraging to see high end attorneys use AI–because it will bring the cost of legal services down. More excitingly is one of the earliest and most important success stories of large language models: AI-based dispute resolution.
This is not a reference to a dispute resolution provider using AI to improve its processes and become more efficient. Such a processes are well underway at the International Centre for Dispute Resolution at American Arbitration Association (“AAA-ICDR”). At the AAA-ICDR, we have been evaluating, piloting and building with AI to improve just about every part of our process from clause drafting to intake to neutral assignment. But these operation upgrades in AAA-ICDR’s current processes are just the beginning.
What comes next is AI-based dispute resolution: two parties to a dispute that decide to resolve their dispute by a neutral AI that “hears” the two sides and delivers a binding decision. This is where AI can have the greatest social impact. Large language models can offer people their day in court at a fraction of the current time and cost, and quite possibly a more neutral decision maker. Ilya Sutskever, the co-founder of OpenAI and now the founder of Supersafe Intelligence said that what surprised him most about the models, if he had to choose one thing was their humanity: “When I speak to the models I feel heard.” This is essentially what disputants need (in addition to a decision) to be heard.
The technology is here. In fact, we can see its effects, a French call center company, Teleperformance, lost significant value after fintech company Klarna announced it was using AI to do the job of 700 customer support agents. AI call center startups Decagon, uJet, Level AI, PolyAI and Parloa have each raised over a hundred million in VC funding. Depending on the specific vertical, between 10-40% of call center interactions involve dispute resolution. Of course, dispute resolution in the context of the call center is not the same as many commercial arbitration or mediation matters, but not every dispute needs those processes we now use. We have the opportunity to fit the forum to the fuss. And across many metrics there is a narrowing gap between human and AI-driven neutrals.
That the technology is ready does not mean people are ready to embrace it. History offers countless examples where human adoption and acceptance lagged behind technological advancements. It took a decade after the invention of the elevator for people to stop taking the stairs. The internet was around for 20 years before users trusted it enough to enter their credit card details. And even today, majority may not feel truly safe in a fully self-driving car.
At the AAA-ICDR, we are laser-focused on ensuring that our AI not only works but is tailored for human use. To us, this means not simply replicating human dispute resolution processes with AI but re-imagining a world with more efficient and accessible resolution channels. This requires extensive testing, raising awareness, and actively listening to feedback from our current users and neutrals. Ultimately, the success of integrating AI to our systems and processes won’t just be measured by how well we serve our existing customers, but by ultimately by our reach and the acceptance by the large part of users of using AI in dispute resolution.
We look forward to attending the SCCA 25 Conference!
________________________
To make sure you do not miss out on regular updates from the Kluwer Arbitration Blog, please subscribe here. To submit a proposal for a blog post, please consult our Editorial Guidelines.
