Hello /sci/, I'm a dummy and don't belong on this board

>>8148740
You do not need a good program, just one that is better than the current system. Turns out computer are bad at making evaluations regarding people behaviour, so I will take a while, maybe forever, to make a good judge computer

>>8148740
>genius idea
>using computer for everything
Yeah, truly breaking news.

>>8148752
I only said that in contrast with calling myself a dummy anon

honestly I just wanted to see if you'd all debate the idea of it's coherent flaws and pluses

please weigh in

>>8148740

Not enough data.

Once surveillance is more widespread and commonplace, and proper pattern recognition neural nets developed, computers can determine whether a crime was committed or not, so long as it has access to the relevant data.
Computers could also be used to determine if someone is lying with 80-90% accuracy for non psychopaths (the accuracy is less than half that with psychopaths). It could be expanded with a large enough database and frequent mandatory testing of the population (like a census sort of thing, just to keep the machine calibrated with usual individual responses to questions the machine already knows the truth value of the answer).
I don't see the justice system being run by machine's anytime soon though.

>>8148740

I think looking at the causes and definitions of crime is where we should be focussing our attention, rather than on how to live with it.

>>8148761
Thanks for the reply anon that was just what I was looking for

For the psychopath part of your post could you link me to something to read about how hard they are to read? for the story part on my end. Thanks again

>>8148779
Yeah but if we were to implement some sort of system how do you think it be done about?

>>8148740
The rule of law is based on three things:
1. Equality
2. Necessity
3. Legal certainty

If you favor 1 over 2 and 3, you'll get natural law and chopping of body parts.
If you favor 2 over 1 and 3, you'll get a police state and 1984.
If you favor 3 over 1 and 2 you'll get legal positivism and a free and happy society.

So in order to have legal certainty, everybody would have to completely understand your computer program. Otherwise people wouldn't be able to regulate their conduct. This would only work in a society of computer scientists, or if your program is so simple, that it isn't necessary in the first place.

>>8148786

>how do you think it be done about?

lol wut?

Well, to be perfectly honest, I really don't know.

I think we'd need to rewire the brain or perhaps inhibit and manipulate particular firing patterns, in relation to psychological mechanisms corresponding to intra-sex rivalry, homicide, kin detection and altruism, just to name a few.

Ant societies are peaceful (when they're not at war with unrelated ants) and efficient as fuck, however that is because all the workers are female siblings that share approximately 75% of their genes with one another.

We humans on the other hand are a lot like our chimp and bonobo relatives, when it comes to social interaction.

>>8148796
So like a since pre kindergarten lam program would need to be implemented? that way even holding child killer accountable. If that was
what you meant then thanks for the reply anywho

>>8148803
I'm sure you'll work it out

>>8148846

Lol, I'll try.

>>8148846
>So like a since pre kindergarten lam program would need to be implemented? that way even holding child killer accountable.

This doesn't have to apply to children. The thing is that a normal, healthy adult should at any point be able to know a legal solution to his problems.

On second thought your program would also be possible if you had always direct feedback from it. Like a computer chip that tells you with certainty whether you'll get punished for something you're about to do.

>>8148865

What about situations where the punishment is less severe than the benefit?

For example, pharmaceutical companies failing to publish trial data on a drug with potential health risk, identified by practicing physicians, being fined $10,000 a day while making billions per year from said drug?

In situations like these, there is no incentive for obeying the law.

Wouldn’t such a program facilitate crime, in such a scenario?

File: pass.jpg (36KB, 225x318px) Image search: [Google]

36KB, 225x318px

A justice system based on computers...well that sure does ring a bell

>>8148930
thanks anon, i\ll check that out

how about your weigh in on how it should or shouldn't be done? it would be appreciated

File: idontknow.jpg (22KB, 600x315px) Image search: [Google]

22KB, 600x315px

>>8148939
As 8148744 said, the current problem would be getting a computer smart enough to understand human behavior, ethics and shit like that. The closest to what you're describing that can be achieved would be probably a point system, where human pre-score what the severity of the punishement is and the computer program just hashes it out- not great since human bias is still there. Now don't hate me, but if the issue was making multiple judgements at the same time was a problem then would'nt quantum computing would solve it ?
I guess the main issue would be how little we understand about ourselves, including our behaviour.

Thats not a genius idea, its a thought.

>>8148740
Algorithms can be biased.
One example:
>https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
Think of it like this: assume that poverty leads to greater likelihood of crime (this is true), now assume that due to racism blacks are poorer than whites (maybe true, I don't know) then a machine would punish blacks at a greater rate than whites "for the same crimes" (italize that, i know that blacks commit more crimes, the point is that you can control for that). The real situation is more complicated and we may not know the specific reasons behind bias, but we know it exists, we can measure it.

Thanks for all the replies /sci/ I wish I could be like you and join but you'll gave me insperation I won't forget. If my novel makes money i'll be back and peace you off, thanks.

>>8148982
appreciated, i'm sorry but I'll read this tomorrow, thanks again.

>>8148740
You think nobody's thought of this before? There are so many problems with this. What if the programmer was somebody from /pol/? What if moral values change (which they will)? What if the code has bugs? Most obviously of all: what if it's wrong? The only way to fix any of these problems is through really intense programming which definitely won't last , as well as removing the absolutism of the machine's decision, which basically ruins the whole fucken point anyway.

>>8148994
No I thought of those, do I have an answer? no I was hoping /sci/ would.. I never did intend on selling my idea I just wanted to see how it would be discussed amoungst smarter minds.. So please anon tell me what youthink

>>8149009
I just did

>>8149013
thanks even though it went right over

i\ll include you in those notes

>>8148740
Nobody from Singapore here??
In that city you can pay your fines before being caught, Singapore style efficiency and probably the closest we are to your vision.

>>8149016
can you link me to me how your city is? please share my man

>>8149018
SG is not my city. I have never been there but William Gibson made a famous article about it: http://www.wired.com/1993/04/gibson-2/
This article got Wired banned there.

See also: https://www.statecourts.gov.sg/CriminalCase/Pages/HowToPayTrafficFines.aspx

>>8148740
No, because computers wouldn't be programmed with jury nullification.

>>8148740
In patent cases you will normally have to do claims construction and Markman hearings. Even though the claims can be viewed as a program there is usually conflicts between the parties as to what the claims should be construed as.

I am not sure computers are anywhere near doing a claims construction.

>>8148994
>What if the programmer was somebody from /pol/?
Society would improve

>>8149026
thank you anons

\!?

thanks again

>>8148761
>frequent mandatory testing
nigga i dont want to live in Psycho Pass

>>8149286
thanks my man

>>8149105
>pol
>improve
now thats edgy

related

https://www.youtube.com/watch?v=CMbKalNq2gs