A computer-driven system for the policing of the world’s biggest cities would be a major technological advance.
Key points:The system could be a powerful deterrent against crime, and could be used for everything from tracking missing children to tracking drug-dealing gangsThe system would be used in real-time to identify threats and then act, as police officers work out how to respondThe system is thought to have a price tag of $1bn to $2bnThe system, called the AI Police Unit, is based on the work of Professor John Kagan from the University of Sydney.
It is said to be one of the most powerful AI systems ever built, capable of processing data from more than a million data sources, including police, social media, maps and video feeds, all of which can be used to predict and intervene in criminal events.
The system will be the subject of a big research project at the University’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and will be developed by Professor Kagan.
It could potentially provide police with a powerful tool, but its biggest potential is in the field of predictive policing.
Professor Kagan told the ABC it would be important to consider how a system would interact with the environment.
“We don’t want the AI to think about how it’s interacting with other things, like cars and people,” he said.
“If you’re driving, the AI will want to know where you are.
It may need to think twice about how you’re interacting with people.”
So, it will be important in the design of the system that it be able to use its environment to make decisions about how to interact with people, and to see what the people are doing.
“You’ll have a police officer that is constantly thinking about what they’re doing and how they’re interacting and how to make the right decisions, and so on.”
The question is, how do you do that?
“The system’s developers say it could help police with the following tasks:The first task would be to determine the location of a person, as a potential crime target, and then the system could use that information to decide how to react.
In the second task, the system would use the information it gathered to determine whether or not to arrest that person, or to take a further step to investigate further.”
It’s a pretty big decision, and it will depend on what kind of information is available to it,” Professor Kagansays.”
What’s the location that we have to get there?
Is there a crime scene, a witness or someone we have no reason to suspect of having committed a crime?
“The AI can make those decisions on its own, but the AI has to be aware of what it’s doing.”
The system then could determine whether to follow the person down the street, or take a more aggressive route, and if the police have reason to believe the person is a criminal, the police could intervene.
“As it’s making those decisions, it’ll also have the information to know if there are other potential crime scenes that need to be investigated further, and when it thinks that it’s safe to intervene,” he says.
“And it will have the right resources at its disposal to do that.”
Professor Kagan believes it would work better if it had to make these decisions in real time, with the police officers working out how best to react to the information they were receiving.
“Imagine that a car is driving in the intersection and it’s going in the opposite direction,” he explained.
“In the last few seconds of that car, the officer is able to see that the driver has been involved in a collision and that the other vehicle is a car, and that car’s driver has escaped injury.”
When they see that, they will have a better chance of making the right decision.
“Now, what if the car that the officer has been following has not been involved?
Well, it’s not clear what the circumstances of that incident are.
So, the decision will have to be made in realtime.”
That’s how a lot of these things work in real life.
So it’s an interesting thing to have that ability to make those kinds of decisions in the context of the real-world, and we have a lot to learn from it.
“The AI Police unit will be based in the Police Technology Centre (PTC), which has been described as Australia’s largest research and development centre for AI and computer science.
Professor Michael Mearns, the director of the Center for Information Science and Innovation, said the system’s capabilities could be as important as the technology itself.”
Computer scientists and computer scientists are all about making technology better,” he told the BBC.”
But the problem is that most of the time it’s hard to get to the next level of computing capability.
“I think what’s really exciting about the AI police is the potential for