Can we really trust legal AI?
For a long time, the law was considered immune to tech disruption. Some great leather-bound tome of human knowledge, gated and sealed off, zealously guarded by big firms and a tangled web of regulatory tape.
“I think lawyers tend to think of the legal process as a complex black box that can’t be unpicked,” says Joel Tito, an AI expert for the Centre of Public Impact in London. “If we’re being cynical, the profession has a vested commercial interest in conveying complexity. They mediate this knowledge and they charge a fee for it. But when you tug on that thread, and you point out the processes are actually quite simple, it undermines a cartel they’ve set up.”
In 2019, it’s fair to say, the thread has been tugged. LegalTech is now a lucrative Silicon Valley industry, raking in $1.63 billion in seed funding last year (up 713% from 2017). Legal technology start-ups are popping up everywhere, from CaseCrunch, which proved algorithms can beat humans when predicting PPI claim outcomes, to Hello Divorce, an AI-powered platform designed to undercut US divorce attorneys (founder and lawyer, Erin Levine, estimates that Hello Divorce can save clients up to $16,000 in legal fees). There are 93 legal tech firms in Australia alone, building everything from data analytics and cloud security to Blockchain-based legal chatbots. Education is also jumping on the train, with RMIT Online recently launching a Graduate Certificate in Emerging Technologies and Law to help legal professionals stay abreast of the changing industry.
But is the Legal AI hype really justified? And where does this leave cigar-chewing human lawyers? How much faith are we willing to place in algorithms, especially when sensitive information and millions of dollars are at stake?
“It’s human nature to be curious, especially about technology and how that might change the way we do things,” says Joni Pirovich from independent firm, Hall & Wilcox,” But it’s also human nature to doubt something like untested and unproven technology.”
Pirovich says the legal industry is at a curious tipping point. On the one hand, firms are worried about been squeezed out by nimble, tech-driven start-ups; On the other, clients and lawyers are cautious about AI’s realistic potential, especially in the wake of high-profile security breaches and tightening data privacy laws. “Our laws reflect this mentality,” says Pirovich. “Like autonomous vehicles, I think we will only trust technology to a certain point, and after that point a human lawyer would review the output and provide feedback.”
This is a growing headache for LegalTech start-ups. Machine learning algorithms can be incredibly efficient, but they rely on clean, readily-available data. The more accurate and complete the data, the better the algorithmic output.
But global data collection is moving in a very different direction: last year the EU introduced the Global Data Protection Regulation (GDPR), which set off a string of similar privacy legislation (including the California Consumer Privacy Act and upcoming reforms in the UAE). Tighter regulations are good for citizens, but bad for tech firms. Jackbooted regulatory investigations and new private rights of action both represent significant risk for investors. And not the fun kind.
Pirovich says the answer might lie in cooperation, rather than Free Market rivalry; a concept with which law firms are generally unfamiliar. “For truly disruptive applications of tech in the Australian legal sector, we need more than just investment – we need law firms to work together,” she says. “We need them to share their data in a legally compliant form. One of the biggest opportunities for the use of AI in Australian legal tech is access to multiple law firms’ data.” Unfortunately, Pirovich admits that such cooperation, given client confidentiality and security concerns, is almost impossible.
Of course, none of this represents a nail in the LegalTech coffin. More of a speed bump on the Progress Expressway. Most experts agree that the big question in 2020 won’t be, “Is legal technology beneficial?” but rather, “Which AI solutions are best suited to which problems?”
Joel Tito says the future will come down to two factors: what the profession is willing to stomach, and how much the public is willing to trust. “People don’t mind having their tax return prepared by an algorithm, but people might living in a cell for six months, based on an AI decision,” he says. “AI is sexy in the private sector, because it reduces cost, but that’s not what excites me. What excites me is improving access to legal knowledge: we need to move the public discourse into the realm of public outcomes, improved outcomes.”
This article was provided by ALTA’s Education Partner, RMIT Online. If you are interested in keeping up with LegalTech trends, find out more about RMIT Online’s new Graduate Certificate in Emerging Technologies and Law, starting on October 7, 2019. The first postgraduate accreditation of its kind in Australia, this program empowers legal professionals to deal with the radical change of digital disruption in the law and policy.