Utilising AI as a decision making tool within the criminal justice system is unethical

An approximation shows that every year, up to 10,000 people are wrongly convicted in criminal justice systems across the globe. Although AI can be implemented in various areas of the legal system to solve these problems and help the work force, there are several obvious flaws in the way that it operates and it could even contribute to even higher rates of unjust convictions.

One such flaw that comes to mind is the use of criminal risk assessment algorithms. These algorithms function by processing the data of a defendant’s profile and generating a recidivism score between 1 and 500. This impacts the Judge’s decision as it is a score on how likely the defendant is to re-offend. However, the defendant may receive an unfair score as the algorithm is based on data which may not be accurate or reliable. Eric Loomis was sentenced to six years in prison due to the high AI system’s recidivism score that he received. An investigative report by ProPublica later revealed that these algorithms tend to reinforce racial bias in law enforcement data

Another issue with the utilisation of Artificial Intelligence in the criminal justice system is that it can be tampered with to change the outcome. This can result in unfair sentences as the algorithm does not perform correctly. It can be tampered to be racially unfair or to look at data in the wrong way.

Furthermore, human judges have the capacity to empathise with the defendant which can cause a more realistic sentence to be implemented. On the other hand, artificial intelligence is incapable of empathy and this could result in cases with more emotionally dependant cases.

Finally, Artificial Intelligence may become capable of rewriting its own code, resulting in an AI system that has been designed by an AI system, this could lead to problems as the AI system may not create a just algorithm for the cases.

Do you want your job taken by a machine?

Artificial Intelligence is impacting almost every part of the industry as well our lives.

Some examples of this are; how we go about the aspects of our daily lives, including transportation, home services, medicine, education, public safety, the job market, entertainment and much more.

Do we want this though? Would you rather a pile of metal and some code do your job? Do you want robots stealing your money? I certainly don’t. It is predicted that within the next decade, AI will be able to write a school essay and drive a truck better than a human can. As well as this AI is said to have a 50% chance of outperforming all human tasks within the next 45 years and automate all jobs in the next century. The question is are you willing to give up your jobs to these machines?

Some effects of Artificial Intelligence taking over jobs

Lawyers – While some legal work may still requires humans, A.I. can help perform tasks ranging from legal discovery to creating contacts . They can even argue about parking fines and can handle divorce cases.

Journalists – Can use bots to generate sports reports ( or any other news articles) or use AI for in-depth investigative journalism.

Chefs – One example of a chef robot was created using IBM’s Watson technology. Called Chef Watson , it’s able to generate entirely new recipes from scratch using an astonishing knowledge of taste chemistry and flavor pairings.

Medics – Whether it’s algorithms which will be able to diagnose diseases, computers being used to make recommendations about the best kind of treatment, A.I. pharmacists, wearable devices that can help treat physical disorders, or even robots who will be able to carry out surgery, there’s no doubt that cutting edge technology will have a big impact on a range of medical professions.

Financial analysts – Like being a lawyer, working in the finance sector has traditionally been a high status, high income job. However, increasingly A.I. is taking over. Computers can spot patterns and make trades faster than even the most eagle-eyed of human analysts. With billions of dollars (or more) at stake, it’s no wonder that machine learning tools are all the rage, while some estimates suggest that around 30% of banking sector jobs will be lost to A.I. within the next decade.

Consequences of AI taking over Industries

A two-year study from McKinsey Global Institute suggests that by 2030, intelligent agents and robots could eliminate as much as 30 percent of the world’s human labour.

Roughly 36 million Americans hold jobs with “high exposure” to automation — meaning at least 70 percent of their tasks could soon be performed by machines using current technology.

AI is different because it can be applied to virtually any industry. When you develop AI that can understand language, recognise patterns, and problem solve, disruption isn’t contained. Imagine creating an AI that can diagnose disease and handle medications, address lawsuits, and write articles like this one.

The transition could be very painful. It’s no secret that rising unemployment has a negative impact on society; less volunteerism, higher crime, and drug abuse are all correlated. A period of high unemployment, in which tens of millions of people are incapable of getting a job because they simply don’t have the necessary skills, will be our reality if we don’t adequately prepare.

McKinsey reckons that, depending upon various adoption scenarios, automation will displace between 400 and 800 million jobs by 2030, requiring as many as 375 million people to switch job categories entirely. 

Using AI to target ads is a price worth paying for free services in social media

Artificial intelligence will allow marketers to recognise objects in consumers social posts to better target ads suited to them and uncover hidden patterns a wider range of data sets. This can further boost a region’s economy rate due to an increase in interest in a certain product.

fig. 1

Targeted advertising allows for more relevant adverts for everyone and increases the profitability for companies as they will not waste money by showing their adverts to people that might not be interested.

Also, not just in advertising but in other aspects of the internet such as video recommendations. If you had been watching many hours of cat videos, it would be expected that the site recommends similar videos as it is evident the you are interested in this topic. This is beneficial for both the viewer and the company because it will make the viewer satisfied with their experience and the company will have exposed itself to the viewer.

fig. 2

In addition, this boosts local businesses which enhances a countries economy

AI will grow to $7.3 billion a year by 2022, up from $2 billion in 2018. This is because companies will invest heavily in AI tools that will help them differentiate and improve the services they offer customers.

Using AI in CCTV is invasive

Imagine being able to see anyone you wanted around the world at any given moment. The era of anonymity will come to an end; AI allows you do to this, without your subject ever knowing. A nation under CCTV

Examples of AI used around the world given by MIT:

Police departments in the US use predictive algorithms to plan where to send their staff during patrols, but the AI tend to focus on less problematic, but more ethnically diverse areas. The AI used repeatedly demonstrates that face recognition systems can fail spectacularly, leading to discriminatory outcomes.

Japan uses open-source AI technology to identify shoplifters. However, the AI frequently mistakes indecisive shopping with suspicious behaviour and reports innocent shoppers to the authorities

Judges are making increasingly incorrect decisions regarding sentencing and parole, relying on stats and predictions made by AI; instead of going with their gut or using their years of experience and education.

Constant surveillance in China has been a cause of concern around the world for many years. This AI not only tracks China’s citizen’s 24/7 but also categorises them based on a “social credit score” which assigns them a social rank in society based on activities that displease the government, such as excessive spending. The system is meant to be 90% accurate and is supposed to increase national safety. There are meant to be over 600 million cameras by 2020

Negative Effects:

Worrying– people may feel unsafe and on edge with someone constantly watching them.

Restrictive– social ranking limits opportunities and can be used against people in the future.

Not that accurate– there is still a chance that mistakes can be made, this could cost and ruin peoples’ lives.

Driver-less Cars, Just Another Way To Get People Killed On The Road?

There have been numerous incidents where AI’s have malfunctioned when doing their task, what do you think happens when the AI driving your car decides to pack up? Lets just say the consequences could be fatal. Basically all software’s can glitch every now and then, but when peoples lives are at stake, there’s no room for error.

Although software may be able to predict collisions in the future, there is no moral decision in who dies and dies when it is entirely calculated by lines of code. People are much better at deciding to either hit the car in front and possibly kill themselves or run over a group of pedestrians crossing the road.

Everyone knows how professional hackers can take over software at will, at any point someone could take control of an automated car and steer it into a ditch, killing everyone inside. Anyone in one of these cars could be abducted by being taken to an unwanted location.

On the 22nd of January 2018, a Tesla model S ploughed into the rear of a fire engine at 65 mph due to the vehicle being on autopilot. Luckily there were no injuries, but who would have been to blame if there were?

Automotive vehicles give potential terrorists another way to send an explosive to any location without being hurt themselves. All they have to do is purchase one of these cars and set a location on the satnav.

AI wont always be able to understand all human signals or intensions. For example, if someone is signalling to cross the road then proceed to as the car is passing without knowledge of the person’s intensions, they could be ploughed through or hit while the car is hitting the breaks, potentially killing them.

AI in social media?

Many people are unaware of how big media companies- such as Facebook and Google- use user inputted data to enhance their own profit. For this, they use Artificial Intelligence (AI).

Is using AI in social media a breach of privacy?

How do Social Media companies use AI to their advantage?

An example of a company that is thriving with the use of AI is Facebook. With over 800 million users logging on to Facebook everyday, the company has many opportunities to collect and use their personal data. But where is this data used? They collect data from recent posts and comments to show users ads that relate to their recent online activity. For example, pages you like and comments.

If you have ever been tagged in a post, your face is now being used as data for Facebook as well as other companies. This is thanks to social medias Facial Recognition software. This gives them the ability to:

• Track your location
• Learn about your relationships
• Find out what you like (hobbies, food etc.)

and more. This gives companies the power to breach your privacy. As a result you may find that your ads are more “catered” to your recent online activity.

Using AI in social media is a breach of privacy.

Why using AI to automate industries makes them more efficient and is helpful for the economy

• 

Imagine this: a human and a machine have been instructed to carry out one task, to fold 1,000,000 paper cranes in a day.

Can you see what might be the problem if it were up to the human to complete this task.

The record for the number of cranes folded in an hour is 20 so even if he went folding these cranes for 24 hour straight he would only be able to fold 480 cranes and that only would happen if the guy didn’t eat or sleep etc.

However with the robot, we could program it to fold cranes very quickly and efficiently and as a bonus, it would complete the job to a constant standard without getting distracted or creating human errors.

There is obviously the problem with human error in the coding for the robot itself,however once it is found, it can be easily changed and so is more efficient than a human because if we found a man making the same mistake, we would probably need to retrain the person.

A company called Ubenwa intended to create a machine which would be a digital diagnostic tool for birth asphyxia.

The team at Ubenwa have managed to turn an expensive, resource intensive diagnostic process into an accurate, cost-effective and lifesaving alternative.

Another obvious disadvantage is that people will become unemployed however this can easily be countered by opening new jobs in coding and developing the AI in question. Many businesses will also not be able to afford AI so these areas of jobs will still be able to employ humans and with the ever growing upstart of new businesses, there will always be new job opportunities for people.

There are hardly any cons to using AI in industry because in the end, they open up new opportunities as well as making industrial processes more efficient overall.

Driverless cars – A case study

DThe roads we drive on are very dangerous, with people making news often, with crashes and deaths. Humans are, so it seems, fantastically flawed creatures. And yet many would trust a human to drive, rather than a computer.

Arguments against driverless cars:

1. They won’t work until cars are as smart as humans. Self-Driving cars could be very well executed, because, ideally, all cars could interface, removing traffic jams altogether, and severely reducing crashes. This would require infrastructure to be in place.
2. Self driving cars will cause jams. Well yes, but actually no. As mentioned previously, all cars could interface and prevent this.
3. They could be hacked to cause problems, like in 9/11. This could happen with any car already, anyone could get in a car and run people over. But security would be high, and death rates might actually drop.

If people are tired they are unable to concentrate and it is known that when you are sleep deprived, there is a decrease in brain function, primarily in one’s alertness and attention, and both are needed for safe driving. Driverless cars would eradicate this issue and ensure the safety of lives. In the US, according to the National Highway Traffic Safety Administration, there are over 100,000 crashes due to people driving whilst being sleep deprived resulting in over 6,500 deaths per year.

Alcohol can have the same effect in reducing the driver’s alertness and reaction time which can prove fatal on the roads. On average in the UK there are 3,500 people who are seriously injured as a result of drink driving every year. The reason alcohol increases accidents is that it increases human reaction time. A self-driving car has a near  reaction time, and as a result, would prevent these issues.

So reasons why a person may need a driverless car:

• Too tired to drive
• Disabled
• Potential distractions? E.g. children
• Drunk

Driverless cars could help eliminate the issue of drunk driving.

Driverless cars could help eliminate accidents due to texting and cell phone use, as well as distracted driving accidents.

It is estimated that driverless cars will result in less braking and accelerating and therefore less gridlock on highways.

If cars become driverless there will be an estimated 80% less emissions, 90% less accidents, as much as 40% less traffic congestion, and space saved in parking.Driverless cars effect on society:According to the US Department of Transportation’s National Motor Vehicle Crash Causation Survey, 94 percent of road accidents are caused by human error, and it is said that driverless technology will drastically lower, if not eliminate this factor.It is predicted that people of 2030 will have some 320,000 jobs in the self-driving car industry open to them, and there will be 2,500 more people alive than there might have been otherwise.A recent study carried out by the SMMT entitled Connected and Autonomous Vehicles: The UK Economic Opportunity, projected that the number of accidents will fall by 25,000 a year over the next 16 years.Also here’s a cool thing- the human reaction time is 0.19 secs, while a computer reaction time is a billionth of a second.

How driverless cars benefit the environment:

– Driverless cars use significantly less gas and fuel, compared to a vehicle driven by a human. Most gas is burned when driving at high speed, braking and re-accelerating excessively. Driverless cars cut these factors out, meaning less gas is burned, resulting in less air pollution.

– Driverless cars mean that there will be fewer cars per household. Due to its high intelligence, these cars could drop the kids off to school, take the adults to work and then park itself until it’s time to pick them back up. Given that vehicles are a significant contributor to global warming, the reduction of cars on the road results in less carbon dioxide being released into the atmosphere.

– They will also reduce traffic which will reduce pollution in more urban areas and allow the authorities to focus on pedestrians and green space rather than traffic flows.

Sources:

Arguments against automated cars: https://www.theatlantic.com/technology/archive/2018/12/7-arguments-against-the-autonomous-vehicle-utopia/578638/

https://greenerideal.com/news/vehicles/driverless-cars-environmental-benefits/

Drink driving stats:

https://www.drinkdriving.org/drink_driving_statistics_uk.php

Sleep deprivation stats:

https://en.wikipedia.org/wiki/Sleep-deprived_driving

Artificial Intelligence : Use in criminal justice system is good

The first advantage and the most important one is the removal of human bias .

Despite all attempts to remove subjectivity, humans will always be emotionally opinionated. People can be emotionally manipulated, thus AI will eliminate the bias as computers cannot be emotionally manipulated and therefore making judgements based on facts alone.

BIAS : An inclination or prejudice for or against one person or group, especially in a way considered to be unfair

• Time saving
• It makes an accurate, swift and just decision, saves time and pressure on a particularly difficult or consequential case. It remove the time used on opinion-making.
• Consistency
• AI could and would give out the same results and punishments on different days consistently, whereas human judges have been known to give different punishments for identical crimes due to their own emotional and mental state.

AI has no emotion towards the decision making, or no guilt for imposing a harsh pronouncement on a person : it is immune to emotional manipulation.

• It makes an accurate, swift and just decision, saves time and pressure on a particularly difficult or consequential case.
• Although nobody likes to act as a jury, AI eliminates the dislike and the hate for the job doing and will be able to be accessed anywhere without the delay from the judge being late
• There are no complications! The AI system looks at the points and judges the value of both cases, and only looks at the valid information and makes the right, fair decision.
• There are no complications! The AI system looks at the points and judges the value of both cases, whether they are valid or invalid and makes the right, fair decision.
• It eradicates the halo effect, which is a type of immediate judgement discrepancy, or cognitive bias, where a person making an initial assessment of another person, place, or thing will assume ambiguous information based upon concrete information, so that data will not be biased based on how they look as a computer feels no emotion

AI within CCTV: Increasing public safety, or an invasion of privacy?

Face recognition: CCTV cameras could have the potential to track you anywhere.

In public surveillance cameras good or bad debate, the first question is whether the surveillance cameras an invasion of privacy.

What is CCTV?

CCTV is used for a number of monitoring and surveillance purposes, but is mainly used for security purposes. CCTV involves the use of an unmanned, remotely mounted video camera system, transmitting live pictures back to a television screen where developments can be monitored and recorded.   

• 42% say yes and 58% say no; according to debate.org in regards to the security cameras and privacy discussion.
• Graph shows the majority of the public do not see the use of cameras as an invasion of privacy.

Advantages

The public security cameras can be used as a way to keep an eye out for any crimes that are being committed. Security cameras are put in public spaces, like a pedestrian crossing, a supermarket, parking area, museums, concerts, do a good job in providing 360° surveillance.

Crimes can be deterred before they even begin, as tracking faces can make it easier to catch criminals and identify them with little difficulty, before crimes occur, therefore with less risk involved.

If a suspicious individual or items are seen in an area, the appropriate authorities can be contacted to take action before any damage is done or any crime is committed. Moreover, any people in the area can be cleared as a precaution.

Since public safety is guaranteed, people’s fear of crime can also be reduced, while their sense of security is created.

Video surveillance statistics show that camera surveillance does reduce crime. According to it, crime incidents in Humboldt Park dropped 20% after the cameras were put in public.

Facial recognition software has improved greatly over the years, which means that if a person is caught on camera committing a crime, the chances of them being caught is much higher.

In some cases the public video surveillance footage can also help prove someone’s innocence if they have been accused of a crime that they did not commit. Human bias and prejudice is eliminated with the use of AI. For example, racial disparities within the police force can be avoided and can lead to a lower number of wrongful shootings.

For example:
In 1984, then 19-year-old Darryl Hunt was arrested and convicted of the rape and murder of Deborah Sykes on the outskirts of Winston-Salem, NC. Hunt’s girlfriend provided an alibi, but when she was arrested on outstanding larceny charges, she later told police that he had confessed to killing her. Despite recanting this false information before the trial, the prosecution used it anyway, ultimately securing a conviction.

Darryl Hunt

The use of AI through CCTV cameras could have aided in avoiding this wrongful conviction of the murder of Deborah Sykes on the outskirts of Winston-Salem, NC. Hunt’s girlfriend provided an alibi, but when she was arrested on outstanding larceny charges, she later told police that he had confessed to killing her. Despite recanting this false information before the trial, the prosecution used it anyway, ultimately securing a conviction.

How can AI improve CCTV?

Artificial intelligence can be used to sharpen low quality images to enable security personnel to draw meaningful information out of them. AI cameras mean better algorithms and It also helps make good use of images and videos that would otherwise be blurred to make for a useful analysis through the naked eye.