AI is Unlikely to Replace Lawyers, Here’s Why

Letter of the Law

Let’s start with the obvious area that AI should be able to excel at over humans. The “Letter” of the Law. There’s something to be said about how in law there is always a “right” answer, everything in it has a clear definition somewhere within the relevant legislation – usually right at the top – that specifically tells you how to read the information contained within. This allows for laws to be precise bodies of textual information that an AI should be able to easily shift through in order to generate an appropriate answer, right? Well…

The thing is, while an AI could make a fantastic search engine, without a LOT of contextual information it still requires a human to guide that search. And while an AI might be able to quickly locate specific information, it’s still not doing anything that a properly advanced search engine doesn’t already do. Lawyers, however, bring a lot more to the table – in particular humans can be self-correcting. For example, an AI that misinterprets a search query or has an error in its processes that causes it to skip certain legislation, cannot refine that process based on its own information and causes issues that cannot be detected by the AI.

Now one could argue that humans skip pertinent details all the time, or find the wrong paperwork, but a human with a search engine would only need to change the query or run it past a colleague to notice something was wrong. An AI doing the same job would be less able to deviate – or even worse, it might identify something correctly as an error and embed an incorrect solution into its code. This sort of behavior is much easier to correct in a human lawyer who doesn’t need someone to spend hours breaking down an enormous code to find the small points of error that cause the bugs.

Certainly, becoming a lawyer in Australia involves a lot of reading dense piles of information that an AI might be able to shift through with greater speed. But even if we take the best-case scenario for AI, it’s still just only doing the job of an advanced search engine, and there is a lot more to being a lawyer than just that.

Spirit of the Law

Now we get into the real issue. Laws are made by humans. Humans are fallible. Mistakes get made. Look hard enough and you can find spelling errors and grammatical errors in existing legislation that real humans simply gloss over. Why? Because humans understand what the law -intends- and sometimes the grammatical constructs or vocabularies of the law are from forty or fifty years ago and haven’t been updated for modern times.

Unlike AI which could – and does – get trapped on such minutiae, humans have the ability to read what was intended and argue that intent in a courtroom. A document clearly intended to protect the rights of certain citizens in a specific situation being used to protect a previously unspecified group in that exact intended situation? An AI can’t do it, but a lawyer can. Not only that, but a lawyer would be able to present an argument to the court that a particular bit of law -should- cover the unspecified group and may call upon the clear spirit and intent of the law in order to back that up even if the same letter would exclude it.

Processing Speed

It may come as a surprise but in many areas, humans can process information faster than an AI can, especially when it comes to any social or courtroom scenario. Let’s face it, humans are complex creatures, and creating a compelling well-argued case in front of a judge or jury involves a lot more than simply having a body of facts regurgitated at the opposition.

Evoking emotion plays a key part in many trials, such as talking in a manner that your peers, the judge, and the jury each find socially acceptable, managing a distraught or uncontrollable client, or when trying to conduct an on-the-stand investigation and attempting to apply pressure in the right way to reveal an error in the opposition’s case or test to see if a witness alters their statement upon cross-examination in any meaningful way.

Right now we struggle to develop an AI that can respond appropriately in even highly constrained circumstances – imagine trying to program one that could take on an investigative or performance role and then requiring it to abandon or rewrite that programming for every single case. Even if it could do all that, the sheer volume of information an AI would need to shift through is immense, and it might take several seconds – perhaps even minutes – to effectively respond, and if your lawyer took minutes to respond you’d probably want to fire that lawyer.

The Legal Issues of Using AI in a Courtroom – Ever

Of course, one of the biggest issues blocking AI from court is simply that they’re not allowed in court. Not only are devices that could be used to record or transmit court processing barred from the courtroom illegally, but the fact is that AI is in murky water to begin with. There are a host of legal issues surrounding AI regarding how they obtain their information and what the contents of that information should or shouldn’t have, all of which would reasonably prohibit their existence inside a courtroom.

So no, Lawyers will not be replaced by Artificial Intelligence anytime soon. Even if AI could surmount the issues of letter of the law, the spirit of the law, and processing speed, the very devices that AI are required to operate with would prohibit their usage in a court. The most that AI will ever do is increase efficiency in certain elements of legal practice, and really, that’s about all it can do.