There Are No “Robot Lawyers” And There Never Will Be (Except in One Country)
It was a trend a few years back for every other legal tech company to declare that it had created the “world’s first robot lawyer.” I think there was more claims of first than Kim Jong-un at a North Korean Olympics. From search companies, to document assembly companies, to chatbots, to pretty basic forms on a webpage, it seemed there were “first” robot lawyers everywhere.
But, just like the Supreme Leader’s claim of getting 11 holes-in-one on the first time he played golf, there may be a bit of “mere puffery” involved.
The Turing test, or imitation game, is only a test for an analogy to intelligence and not a test for humanity or some aspect thereof. That is, the idea behind the Turing test is that if a machine is able to imitate a human such ghat, it is indistinguishable from other humans, then it should be considered intelligent.
There are chatbot competitions where the programs try and catfish humans, and if the robot is declared to have passed the Turing test and is, therefore, “intelligent.” I think there have been more declarations of “first robot to pass the Turing test” than disease cures invented by the Supreme Leader. But although there was an official Alan Turing, there is no official Turing test, by design or by organisation, and so it is difficult to premature adulation. I do think that Turing test is something that is difficult but reasonably achievable, like Kim Jong-un giving up Emmental cheese binges.
And so too can be have technology that imitates the work of lawyers. From finding an unhelpful precedent, to entering text into someone else’s precedent, to mindlessly repeating the same general advice, to filling in a terribly designed form, there is an expansive list of legal tasks that can, will, and are done by machines.
But in my view, the most essential parts of a lawyer are not replicable by a machine. Let me demonstrate with the design of a program for the World’s First Robot Doctor.
Step 1 – Ask patient how they feel.
Step 2 – Print: “Take an aspirin and come back in 24 hours”.
Step 3 – Ask patient if they are cured.
Step 4 – IF “Yes,” THEN END, IF “No,” THEN GO TO Step 1.
This “Robot Doctor” does imitate the actions of many human doctors and is likely to lead to a recovery by the patient in a large percentage of cases. It could be improved by the incorporation of prayers to Kim Jong-un, as they are reported to have spontaneously cured both AIDS and cancer (but not starvation).
The absurdity of this “Robot Doctor” is obvious because of its simplicity. However, I am confident that I could genuinely fool people if I made it sufficiently complex and using the “latest in machine learning” to derive a similarly useless program.
In my view, the fundamental problem with the “Robot Doctor” (even if it was a more complex program) is its inability to deviate from rules. That is, even though it may be a common “rule” or “system” for doctors to treat patients with low levels of illness with “take an aspirin and call me in the morning” they are not actually bound to follow that rule. They could, at any time, decide that even though they would normally treat a patient using a particular rule, that in this case, they will do something different. And that ability to break the “rules” is how rare or unusual conditions are treated with “lucky” results and advances to medicine pioneered.
In the field of law, there is much of the law and legal practice that can (and will) be systemised and then automated into programs. However, the critical aspect of a lawyer that only a human may possess is that they could at any time choose to deviate from the existing rule set. A lawyer could say that a client’s case is bound to fail under existing laws and precedents but that if new law was created at an appellate Court, then the unfairness against the client could be remedied. There might be some non-legal solution to the problem at hand – there are plenty of cases that are motivated by pride, revenge, jealousy, or anger. Law is more than just the application of and enforcement of rights and obligations.
Regardless of the complexity of the rules that a robot is given, it can never choose to challenge those rules. Even a program to change other parts of a program is adherence to that set of rules. A lawyer is one who is the master of the law (at least in their domain of it), and to have mastery of a subject is to not only understand its benefits but also its flaws. And to have the potential (whether or not exercised) to challenge those flaws. Only a human is capable of such a challenge. That is why there is no such thing as a robot lawyer, and there never will be.
Except, of course, if the brilliant scientific mind of Kim Jong-un was ever applied to it. Because every law and creation that he has ever made has been met with 100% success and approval. Maybe North Korea is the future of legal tech?