Can AI Deal with a Crazy Client?
When court cases are driven by personal motives, logic can be a victim of the fallout. One needs to look no further than the case of Dawnell Batista, who filed for divorce from her husband after 15 years of marriage. In response, her husband, Dr Richard Batista, demanded the return of his kidney as part of the settlement, which he had donated to her in 2001. Failing its return, he pressed for the monetary value of the kidney — which he estimated to sit at US$1.5 million.
In Australia, a neighbour recently spent $80,000 in legal fees to prevent the building of a $50,000 fence. The feud has thus far resulted in 1,140 logged Council incidents over an 18-month period.
Such examples are common, but not rational. This is precisely why they pose a problem for AI in law. With artificial intelligence coming to play a deeper role in legal practice, can client relationships, emotionally-driven claims, and human motives be automated?
What AI Can (and Can’t) Do
From chat bots to driverless cars, AI is staking its claim in consumer industries. However, when it comes to human emotions, AI hits its limits.
“The problem is that law is adversarial, adaptive, and changing,” said McCarthyFinch CEO Nick Whitehouse.
“A driverless car, for example, recognises signs which are roughly the same around the world. There are only so many variables. Because law is not always rational, there can be endless variables.
“A divorce bot, for instance, is great in theory. However, if the other person absolutely hates you, the divorce bot might not be able to compute the kinds of irrational obstacles the other party throws in front of you.”
Reading emotion and discerning context is particularly challenging for AI.
“Prediction of emotions is based on sentiment; you can read natural language and body language, and infer what emotion is being displayed,” said McCarthyFinch’s Legal Tech Counsel and Legal Services Manager Jean Yang.
“However, how do you determine which context is being displayed? How many times have you heard someone say ‘I’m fine’, when it means one of a million different things?
“‘I’m fine’ blows out into a massive, contextual thing. When you take that as a complex problem and apply it to a legal scenario, it becomes even more complex. Lawyers need to determine what their ideal outcome is, and what they’re actually likely to get. That is a challenge for AI entering the law.”
Leave Emotion to the Lawyers, and Data to the Robots
As McCarthyFinch has previously discussed, WebMD provides a sound analogue for the manner in which AI can inform citizens of their legal rights. It also serves as an example of how professionals can use tech to manage client relationships.
“You don’t have to look much further than what doctors have needed to do in the past 15 years,” said Whitehouse.
“They’ve had to compete with Google and WebMD, so there’s been a complete shift in what bedside manner has to be. It’s no longer ‘the doctor knows best’ — it’s now a consultation. ‘What ails you? Why do you think it ails you?’ Bedside manner is now about making the patient feel comfortable and confident that the doctor has a treatment plan in place.
“How many more lives have been saved because people second-guess their doctors?”
AI still has some ground to cover before it can reliably manage client relationships, fully respond to emotions, and contend with adversarial problems. However, it can still be applied to resolve these problems — by pairing up with a human lawyer.
“The solution is to focus on what AI is great at doing — looking at 1,000 cases at once,” said Yang.
“It’s the pragmatic things that are going to improve access to justice; improve lives of lawyers and clients; and improve the legal system.”
“You can develop tech for the sake of developing tech, as academics often do,” added Whitehouse.
“However, when you talk about pragmatic things that help the world, why would you target something that a human can already do exceptionally well? Why not focus on things that AI already does well, and determine how it can contribute to pragmatic points that will make a difference?
“This is why the idea of augmenting the expertise of people is where AI excels. So much of where humans excels is through the expression of emotion, so why should AI target that?”
AI is increasingly crucial to the modern practice of law. However, lawyers will always have a role — whether they’re helping neighbours feud over fences or talking divorcees down from bizarre settlements.
McCarthyFinch is an artificial intelligence platform for automating and innovating the business of law. Founded in 2017, McCarthyFinch is a 50:50 joint venture between MinterEllisonRuddWatts and Goat Ventures.
Follow us on LinkedIn or Twitter, or sign up to our newsletter for project updates, news and latest developments.
