Hacker News new | ask | show | jobs
by eyeJam 3304 days ago
Most large commercial lawsuits settle before they ever go to court. From my anecdotal experience, maybe only 10% go to court. Litigation is a very long and expensive process. Some files last decades before they resolve. When litigation is ongoing, both sides are engaging in legal wrestling to try and get the most favorable position. Once one side has gained a clear advantage, they will start pressuring the other side to settle. Its very expensive to go all the way to court and then lose, so most corporations will concede and settle once they lose the upper-hand. Therefore if a case does finally make it to court, its because (1) both sides have a decent chance of winning, or (2) one side - or both - hate each other so much that they refuse to settle. The first option, (1), is the most common: the case is so close that its essentially a coin flip. Remember, you only hear about the cases that go to court, and most settlements are confidential and never hit the news. So the public's perception of commercial law doesn't include the full picture.

I'm almost positive that litigation will never be fully automated. Some aspects of it already have (such as document review and discovery). But the majority of the work (negotiating with the other side, questioning witnesses, arguing pre-trial motions in court, researching legal questions, writing memos and briefs, advising clients on strategy) is far too reliant on humans to ever be replaced.

3 comments

> (1) both sides have a decent chance of winning, or (2) one side - or both - hate each other so much that they refuse to settle.

There are three other categories of cases you missed and (1) can and should be divided up further into two categories. The third situation arises when one party to the suit simply cannot afford to lose under any circumstances but somehow can still afford to keep paying lawyers indefinitely. Frequently you will see this with large companies fighting to prevent or delay a class action or against an adverse regulatory decision which poses an existential threat to their industry. $50m in legal bills per year doesn't even register against a multi-billion dollar class action risk.

The fourth situation is when one of the parties has a non-financial incentive to continue the litigation, often from a political perspective. For example, when Hobby Lobby in the US sued to get out of the Obamacare birth control coverage requirement. They would have spent themselves into the ground on that case because the principle mattered more to the litigants than the cost.

The fifth situation is when one or both parties have bad lawyers. We (lawyers) don't like to admit how often this happens, but frequently one of the lawyers just completely misinterprets the law or fails to recognize essential facts. They may oversell the case to the client and then paint themselves into a corner which they can't gracefully back out of, and so drag the case out well past its proper expiration date just to avoid admitting they screwed up.

Regarding your (1) point, the evidence for both sides is roughly evenly matched in some cases, so the judge or jury's interpretation of the facts will dictate the outcome. How people weigh and interpret evidence can be incredibly unpredictable, which may prevent settlement in those cases. In other cases, the law is unclear or unsettled, whether because previous cases were poorly decided, a law was poorly drafted, or the issue is truly novel. So unclear law or unclear facts. In an ideal world, those would be the only reasons that anyone would ever go to court.

All of these situations are extremely rare. In terms of quantity of cases, unfortunately people that hate each account for the overwhelming majority, though that tends to be most common where the parties have a B2C or or close business associate relationship, or in non-commercial litigation (most visibly and notably in family law cases). B2B/arms-length commercial relationship litigation tends by and large to be a more coldly rational affair, though not without exception.

You are correct, and I admit that I grossly simplified the situation to make it succinct.
Actually I think the biggest argument against the "profession x will be replaced by machines" reflex is the absence data to train an algorithm. All of the key information is unstructured in emails, decks, meetings and conference calls. How do you train even infinitely better algorithms?

And also what is it going to cost to staff AI professionals to create and support an algorithm vs the salary savings of the 100 guys who must do that job in the UK...

My wife is a commercial litigation lawyer in the UK and 10% is the figure she quoted to someone last night in a discussion.

Also worth noting that, at least here in Scotland, advocates (who are our equivalent of barristers) still get paid for the case even if it settles!