UK Police to start using Minority Report-esque AI to decide whether suspects should be kept in custody

[randomhkkid]

Source: https://www.theverge.com/2017/5/10/15614980/uk-durham-police-ai-risk-assessment-policing

Seem to be on an AI and machine learning kick today, keeping that going UK police are about to utilise a predictive artificial intelligence system that will determine whether a suspect should be kept in custody.

 

"designed to classify individuals based on a low, medium, or high risk of committing a future offense. Police plan to put it live in the next few months to test its effectiveness against cases in which custody specialists do not rely on the system’s judgement."

 

It remains to be seen if this system will have the same inherent bias against certain ethnicities as previous systems.

 

What do you all think about the future of predictive law enforcement?

[vanished]

Once we get computers implanted in our brains, (which will take many years still), by then all privacy laws will have been obliterated and all data stored and transmitted digitally will be basically public, and so the police will be able to simply scan your thoughts in real time and arrest anyone thinking something bad.

[WereCat]

 

Once we get computers implanted in our brains, (which will take many years still), by then all privacy laws will have been obliterated and all data stored and transmitted digitally will be basically public, and so the police will be able to simply scan your thoughts in real time and arrest anyone thinking something bad.

Damn... I would be already in jail for life... best case scenario.

[Creed1]

If technology keeps going like this, we may actually see a real life CTOS (From the game Watch dogs. Its a system that can detect crime)

[randomhkkid]

 

Once we get computers implanted in our brains, (which will take many years still), by then all privacy laws will have been obliterated and all data stored and transmitted digitally will be basically public, and so the police will be able to simply scan your thoughts in real time and arrest anyone thinking something bad.

Isn't the future bright!

[vanished]

 

Damn... I would be already in jail for life... best case scenario.

Most people will would be

[randomhkkid]

 

Most people will would be

 

 

If technology keeps going like this, we may actually see a real life CTOS (From the game Watch dogs. Its a system that can detect crime)

 

 

Damn... I would be already in jail for life... best case scenario.

It's gonna be an immensely interesting future from a philosophical and technological point of view. Laws will have to change to reflect the change in norms and what is considered 'private'. 

[vanished]

 

 

 

It's gonna be an immensely interesting future from a philosophical and technological point of view. Laws will have to change to reflect the change in norms and what is considered 'private'. 

And this isn't even considering the impending loss of every job to robots and AI

[randomhkkid]

 

And this isn't even considering the impending loss of every job to robots and AI

On the upside productivity will go through the roof for creative and lateral thinking tasks as we eliminate the human modality bottleneck of speech/writing. Also language won't be a boundary if we can interpret ECG brain signals in real time.

[WereCat]

 

Most people will would be

Makes me think. If most people would end up in jail... would it still be jail?

The jails would have to be huge and many of them with a ton of security. The people out of jails would be minority trying to keep the other people in jail, resulting in their life style be jail-like as well...

Damn, I don't know how to describe what I want to say in a better way, hopefully it is not such a big mess of words as I think

[randomhkkid]

 

Makes me think. If most people would end up in jail... would it still be jail?

The jails would have to be huge and many of them with a ton of security. The people out of jails would be minority trying to keep the other people in jail, resulting in their life style be jail-like as well...

Damn, I don't know how to describe what I want to say in a better way, hopefully it is not such a big mess of words as I think

Jails are pretty fundamentally broken as they are a huge waste of human potential. I guess jails will evolve/adapt so that inmates do something other than waste tax money.

[samcool55]

 

Makes me think. If most people would end up in jail... would it still be jail?

The jails would have to be huge and many of them with a ton of security. The people out of jails would be minority trying to keep the other people in jail, resulting in their life style be jail-like as well...

Damn, I don't know how to describe what I want to say in a better way, hopefully it is not such a big mess of words as I think

If we are all in jail we would basically be locked up pets with robots as owners.

That's how evolution goes

The smartest thing can control every living thing that's more stupid than it.

[randomhkkid]

 

If we are all in jail we would basically be locked up pets with robots as owners.

That's how evolution goes

The smartest thing can control every living thing that's more stupid than it.

I'm ready to welcome out robot overlords

[WereCat]

 

If we are all in jail we would basically be locked up pets with robots as owners.

That's how evolution goes

The smartest thing can control every living thing that's more stupid than it.

In future people will no longer need to jailbreak iPhones, they get to the next level

 

Depends on how implemented the security will be, the robots may not be more smart than people but they can still be extremely effective.

 

Jails are pretty fundamentally broken as they are a huge waste of human potential. I guess jails will evolve/adapt so that inmates do something other than waste tax money.

Jails right now are in most cases inhuman, instead of trying to correct and improve the life style and behaviour of people they destroy their lives.

There are few exceptions but it is really complicated.

You need to treat every individual differently which is a problem considering how many people are already in jails.

[vanished]

Focusing more tightly on this story now, yeah, I wonder how they're going to do it.  Obviously it will never be able to be 100% accurate, even if it could read your thoughts.  But since we're not even close to that yet, they will have to rely on some sort of big data and correlations, and that will probably lead to extensive stereotyping.  I mean, it basically will be mathematical, systematic stereotyping  

[BuckGup]

 

If technology keeps going like this, we may actually see a real life CTOS (From the game Watch dogs. Its a system that can detect crime)

We already have it bro

[randomhkkid]

 

Focusing more tightly on this story now, yeah, I wonder how they're going to do it.  Obviously it will never be able to be 100% accurate, even if it could read your thoughts.  But since we're not even close to that yet, they will have to rely on some sort of big data and correlations, and that will probably lead to extensive stereotyping.  I mean, it basically will be mathematical, systematic stereotyping  

Yeah, this is why at least in the foreseeable future it will only be an augmentation to human investigators. Basically a system to catch more people slipping through the gaps rather than replacing humans outright.

[samcool55]

 

In future people will no longer need to jailbreak iPhones, they get to the next level

 

Depends on how implemented the security will be, the robots may not be more smart than people but they can still be extremely effective.

 

True, the most important part is that robots have feelings and sympathy like most of us do.

But the thing is that we won't know if that will be the case, that's why the big IT guys say AI is dangerous and should be very controlled. 1 experiment that goes wrong because it breaks rules or becomes too intelligent can ruin everything. We shouldn't be scared of terrorism, but we should be scared from robots, very scared.

[WMGroomAK]

 

Yeah, this is why at least in the foreseeable future it will only be an augmentation to human investigators. Basically a system to catch more people slipping through the gaps rather than replacing humans outright.

Not sure about the UK, but I would think that in the US, someone would try to make a case for police profiling and sue the police department for millions of dollars.  

[randomhkkid]

 

"inherent human bias built in that made them twice as likely to flag black defendants as future criminals and far more likely to treat white defendants as low-risk"

 

So it's racist too.

Racist because it learnt human bias, that's hard to remove when the data it learns from is inherently racist.

[Jito463]

 

"inherent human bias built in that made them twice as likely to flag black defendants as future criminals and far more likely to treat white defendants as low-risk"

 

So it's racist too.

Stereotyping works, though.  And it works, not because of racism or inherent bias, but because we humans operate on the basis of pattern recognition.  Black people tend to commit more crimes (at least in the US, if you break down population versus crimes committed by race, blacks play a far more significant role in crime), which causes them to be considered more likely as suspects in a crime.  It's not racist, it's just recognizing a discernible pattern.

 

As an example, Israel has never had a hijacking on one of their airplanes.  Why? Stereotyping.  It may sound racist or prejudiced, but it's not.  It is biased, but there's nothing inherently wrong with bias.

 

We're typically biased for or against many things throughout the day, it's just that we don't tend to think of them as biases.  I will only use gasoline without ethanol.  That's a bias I have against using ethanol.  Is that bad?  Not at all, but it is a bias.

 

================

 

Anyway, on the topic from the OP; I disagree vehemently with this decision.  We should never leave decisions about the future of an individual up to a computer.  One might make the argument that computers would be more objective, but that argument only holds water if the programmer of the AI was objective.  One could also make the argument that people could be bribed, whereas a computer could not.  That argument falls apart when you realize that a computer can be hacked, while a person could not.  There's no perfect system in life, but this type of judgment should never be left up to an AI.

[ChineseChef]

1 hour ago, Ryan_Vickers said:

Focusing more tightly on this story now, yeah, I wonder how they're going to do it.  Obviously it will never be able to be 100% accurate, even if it could read your thoughts.  But since we're not even close to that yet, they will have to rely on some sort of big data and correlations, and that will probably lead to extensive stereotyping.  I mean, it basically will be mathematical, systematic stereotyping  

Makes one wonder, will the system be called racist/sexist if the AI determines that a specific race or gender are more likely to be a problem?  This group should be held because they cause more issue, that group should be let go because they more reliably show up to court?  If we truly gave the reins to the AI, would the rich no longer be able to buy their way out of jail?  If the AI truly found that a specific group is a problem, how will people react? 

[Centurius]

Okay UK, I've told you this before but I will repeat it again. 1984 was a novel, not a tutorial.

[Taf the Ghost]

The mental gymnastics this is going to cause should be interesting.  The results are pretty straight forward on most of this stuff, but none of it is going to be acceptable to implement. 

[Taf the Ghost]

53 minutes ago, ChineseChef said:

Makes one wonder, will the system be called racist/sexist if the AI determines that a specific race or gender are more likely to be a problem?  This group should be held because they cause more issue, that group should be let go because they more reliably show up to court?  If we truly gave the reins to the AI, would the rich no longer be able to buy their way out of jail?  If the AI truly found that a specific group is a problem, how will people react? 

Considering every time the sub-groups that are a problem are pointed out, they will Kill the Messenger.  That's why the mental gymnastics are going to be hilarious.