Judge decides case based on AI-hallucinated case law

[Paul Farquhar]

Quis custodiet ipsos custodes?

 

[Ryujin]

Actually, the luck of the draw have much more influence on sentencing that we'd like. Studies have shown that sentences are more lenient for people tried after the lunch break than before, when tried for similar circumstances. Even if we remove factors that obviously could influence the outcome (like outright bribery), we still have an irreductible part of randomness. It's quite uncomfortable to think that someone sentenced to 10 years might have been sentenced to 8 years if things totally outside the scope of control had turned differently. But that's the limitation of the system.

When making statements like this keep in mind one simple fact; correlation is not causation.

 

[Jfdlsjfd]

When making statements like this keep in mind one simple fact; correlation is not causation.

The study (from the field of neurosciences) tended to ascribe causation, through a slight human bia linked to tiredness, even when taking other factors into account. The limitations of the study were indeed quite discussed (notably, the timing of hearing case was correlated with the presence or absence of a lawyer, so it was something that had to be taken into account to revise the initial findings). [side note: when it comes to communuication, the initial findings were circulated more than the corrected ones, despite the corrected ones being more exact and more interesting]. What made you think the study was mistaking correlation for causation?

 

[Ryujin]

The study (from the field of neurosciences) tended to ascribe causation, through a slight human bia linked to tiredness, even when taking other factors into account. The limitations of the study were indeed quite discussed (notably, the timing of hearing case was correlated with the presence or absence of a lawyer, so it was something that had to be taken into account to revise the initial findings). [side note: when it comes to communuication, the initial findings were circulated more than the corrected ones, despite the corrected ones being more exact and more interesting]. What made you think the study was mistaking correlation for causation?

The fact that that you linked no studies, so the regimen used is not clear, for one. There are a host of other factors that can affect sentencing, based on timing. A case running longer, because a plea is not taken. A tendency to put the quicker cases earlier in the day so that they can be cleared, leaving time for the more complex cases. All sorts of simple procedural minutia can result in cases that would more naturally result in longer sentences being later in the day.

 

[Jfdlsjfd]

The fact that that you linked no studies, so the regimen used is not clear, for one. There are a host of other factors that can affect sentencing, based on timing. A case running longer, because a plea is not taken. A tendency to put the quicker cases earlier in the day so that they can be cleared, leaving time for the more complex cases. All sorts of simple procedural minutia can result in cases that would more naturally result in longer sentences being later in the day.

Sure, that where the "other factors". Comparisons that are not all things being equal can't lead to meaningful results.

 

[Umbran]

When making statements like this keep in mind one simple fact; correlation is not causation.

Also, if you can point to a cause (like lunch) it isn't random. And our lack of personal control also does not equate to randomness.

 

[trappedslider]

parts of this discussion reminded me of when IBM put urban dictionary into Watson and then had to delete it.

IBM removed the Urban Dictionary from Watson’s memory after it got too sassy

The Verge is about technology and how it makes us feel. Founded in 2011, we offer our audience everything from breaking news to reviews to award-winning features and investigations, on our site, in video, and in podcasts.

www.theverge.com

 

[Jfdlsjfd]

Also, if you can point to a cause (like lunch) it isn't random. And our lack of personal control also does not equate to randomness.

Nice correction. I am sorry for using the word random inappropriately when I meant to say that the time one's case is heard is both unpredictable and outside of the personal control of the tried party and, mostly, the court itself. I thought it was intelligible. I'll make sure to be more precise going further and no longer use "a random accident" when someone is hit by a car or a meteorite while going to a park.

I've removed this side comment as it wasn't pertaining to this discussion and obviously confusingly-worded.

 

[Mustrum_Ridcully]

Is AI intended to give fool proof advice?

It matters more what people perceive it as - if they think it's reasonable, good advice most of the time, but it isn't, that's a problem.
And a knife would probably absolutely be forbidden if people got killed or hurt by them commonly because they accidentally use them wrong and end up not cutting non-human meat but other people. But that doesn't happen with most available knifes, most people can avoid serious accidents most of the time. And those that do harm with them can get into serious trouble.

You might need a knife user license if it requires special training to use them, or only allow people wearing anti-knife gear in the environment they are used. This is the kind of stuff that happened with cars - they are intended to get you from A to B, but accidents happen and get people killed, and it happens often enough that you need a license to operate them and various safety features are now mandatory to ensure that when accidents happen, the consequences are reduced.

So maybe you can use your AI, but you need an AI operator license where you get trained to verify the sources of the AI and find additional sources that could be used to falsify claims made by the AI, or you might need a special operator license to offer AI services, that give you strict legal responsibiltiies if your costumers end up being mislead by the AI or people get hurt.

 

[Ryujin]

It matters more what people perceive it as - if they think it's reasonable, good advice most of the time, but it isn't, that's a problem.
And a knife would probably absolutely be forbidden if people got killed or hurt by them commonly because they accidentally use them wrong and end up not cutting non-human meat but other people. But that doesn't happen with most available knifes, most people can avoid serious accidents most of the time. And those that do harm with them can get into serious trouble.

You might need a knife user license if it requires special training to use them, or only allow people wearing anti-knife gear in the environment they are used. This is the kind of stuff that happened with cars - they are intended to get you from A to B, but accidents happen and get people killed, and it happens often enough that you need a license to operate them and various safety features are now mandatory to ensure that when accidents happen, the consequences are reduced.

So maybe you can use your AI, but you need an AI operator license where you get trained to verify the sources of the AI and find additional sources that could be used to falsify claims made by the AI, or you might need a special operator license to offer AI services, that give you strict legal responsibiltiies if your costumers end up being mislead by the AI or people get hurt.

In many places the following already applies: Carry a butcher knife in your house and it's a cooking utensil. Carry a butcher knife in public and you can be charged for possession of a weapon.

I had police hang around and watch me unload my car when I was carrying a half dozen swords, a couple of bows, a bunch of arrows, and some chainmail for a photoshoot with a friend, at his downtown Toronto photo studio.