“Here Come De Judge”

BBC Article 16 Aug 2021: Would you let a robot lawyer defend you?

In my short story “Apeman” (from the Arkships series), I introduced the idea of robot lawyers:

“Two identical GV34 AI units had been uploaded with Lawyer programming, each with full access to the central database of Case Law. One was designated Prosecution with instructions to argue against the accused, the other was designated Defence for the converse reason. Arguments would be prepared by the advocates: I was the defence advocate and someone else (we don’t know who) would be the prosecution advocate. Only the lawyers were permitted to appear and speak in court (except for the final statement from the accused) and they would use arguments as appropriate based on case law and the direction of the trial.”

Yesterday’s article from the BBC looks at the subject in more detail, including places where Artificial Intelligence (AI) is already being used for legal applications.

In “Apeman“, the lawyer suggests that it is not the job of the court to deal with absolute truth, only to consider whether the defence story or the prosecution story is more compelling. This is in stark contrast to every courtroom drama you will ever see in film or TV.

It’s obvious why the entertainment industries take the line that the court is a Bastion of Truth – our culture depends upon that being the case. Even that wonderful TV series “Rake” fell into line, despite its anarchic theme. The British Legal System (which the “Arkships” series adheres to) was exported to all our colonies and still remains largely untouched in the now independent regions from Australia, through India to the US. Who dares say anything that may be seen as negative against such an edifice, particularly as its proponents are by definition, highly litigious?

Science Fiction, however, has the advantage of looking at our social structures from another point of view. Given certain assumptions, what could possibly go wrong? I think anyone would agree that’s exactly what Apeman is about, although it doesn’t really explore what could go wrong with The Law.

Anyway I’m no expert in matters legal, so I’m not really qualified to comment on that. But I am an expert on AI and the social impact of computer systems – that’s been my focus for most of my working life. And the BBC article reveals some deeply disturbing points regarding the use of AI in law. I refer you to the remarks at the end of the article from Prof Susskind (no, not that Prof. Susskind):

“So if an AI system can very accurately (say with 95% probability) predict the outcome of court decisions, he says that maybe we might start thinking about treating these predictions as binding determinations, especially in countries that have impossibly large backlogs.”

Every business I have ever worked with has specified that their computer system’s success must be measured against Key Performance Indicators (KPI’s). Customer Satisfaction, Profitability, Efficiency, Accuracy are all examples of KPI’s, depending upon business objectives and risks. These companies know that you have to also set up a monitoring system to measure system compliance for the KPI’s to have any meaning.

So if the primary KPI for the British Legal System is not Truth or Compliance, but just the ability to reproduce the current outcomes, are we going in the right direction?