Police in the UK could soon thwart violent crimes before they happen.
This is the aim of a Minority Report-style CCTV surveillance system that is set to use AI to predict whether an individual might be about to commit a crime.
Individuals that are flagged by the system, which is being led by the West Midlands Police, could be given counselling to stop them from carrying out crimes.
The project’s researchers are planning to release a prototype by the end of March next year – but experts say the proposal has ‘serious ethical issues’.
Scroll down for video
Police in the UK could soon thwart violent crimes before they even happen. This is the aim of a Minority Report-style CCTV surveillance system that is set to use AI to predict whether an individual might be about to commit a crime (stock image)
According to an exclusive report by New Scientist, the National Data Analytics Solution (NDAS) system uses AI and statistics to work out the chance of an individual committing gun or knife crime.
It can also work out the chance of someone being a victim to crime or modern slavery.
Iain Donnelly from West Midlands Police who is leading the project says local health and social workers could be alerted.
They could then offer counselling to an individual that has mental health issues which could deter them from committing the crime.
Victims could also be contacted by social services, Mr Donnelly said.
One difficulty, however, is knowing whether the individual’s marked out by the AI system would have actually gone on to offend without intervention .
Sandra Wachter at the Oxford Internet Institute said: ‘How would I know that this actually makes the right decision? That’s something that is very hard to measure.’
Based on a novel by science fiction writer Philip K. Dick, Minority Report, starring Tom Cruise (pictured) is a thriller set in 2054 when police use psychic technology to arrest and convict people before they commit crimes
Researchers on the project have access to the records of five million people who have been stopped and searched or committed crimes in the UK.
The AI is able to find patterns in criminal activity. For example if a number of crimes have been committed in an individual’s social group that increases the chance of them committing a crime themselves.
In total they found 1400 indicators which could be used to ascribe a risk score to each individual on the data base.
According to a report written by the Alan Turing Institute in London that was seen by New Scientist, there are ‘serious ethical issues’ with the proposal.
The report, which is published next week, questions whether it is ethical to intervene if an individual has not yet committed a crime.
The UK is not the only country looking into predictive policing.
WHAT ARE PREDICTIVE POLICING SYSTEMS?
Predictive policing systems can forecast when and where crimes occur using based on prior crime reports and other data.
Palantir Technologies has licensed its predictive policing software with local and international governments.
Most ingest vast amounts of data, including geography, criminal records, the weather and social media records.
From that, it makes predictions about individuals or places that are likely to be involved in a crime, according to the Verge.
There are other predictive policing systems out there that are being utilized, many of them are different.
The Los Angeles Police Department, New York Police Department, Chicago Police Department and, now, the New Orleans Police Department use predictive policing. File photo
Chicago’s police department uses a notorious ‘heat list,’ which is an algorithm-generated list that singles out people who are most likely to be involved in a shooting.
However, many experts have identified issues with Chicago’s heat list.
The government-funded RAND Corporation published a report saying that the heat list wasn’t nearly as effective as a standard wanted list.
It could also encourage a new form of profiling that draws unnecessary police attention to people.
Another academic study found that the heat list can have a ‘disparate impact’ on poor communities of color.
A California startup called PredPol also built predictive policing software that’s been utilized by law enforcement officials, including the LAPD.
In 2016, researchers conducted a study where they reverse engineered PredPol’s algorithm and discovered that it replicated systemic bias against communities of color that were over policed.
It also found that historic data isn’t a good indicator of future criminal activity.
The NYPD also had an agreement with Palantir Technologies to use its predictive policing systems.
An Israeli security and AI research company will soon use AI to analyse the terabytes of data streamed from CCTV cameras in public areas in India.
The partnership has been formed between Tel Aviv-based company Cortica Best Group in India, according to an article by Digital Trends written in April.
The company will analyse ‘behavioural anomalies’ to spot when someone might be about to commit a crime.
This kind of citizen-monitoring technology is already in use in 40 local governments in China.
The technology will monitor individuals by looking for small twitches that might mean they are about to do something illegal.
Co-founder and COO of Cortica, Karina Odinaev, said the technology could identify movements often overlooked by security teams, potentially making cities safer.
She said ‘unsupervised learning’ is required for the software to learn what to spot, which is why they want to train it on security cameras.
Like humans, this technology needs to see something multiple times before it learns to recognise key warning signs.
If the system makes a mistake, programmers can find out which file was responsible for the dodgy call and re-teach it.
The technology could be used for different types of surveillance and could also monitor passenger behaviour using footage obtained from drones and satellites.