Helping identify security threats

Augmented intelligence: Helping identify security threats

It’s a widely known fact that our society is watched more closely these days than ever before.

There are over 600,000 CCTV cameras in London alone, for example. In the USA, there are over 50 million. Being observed goes beyond cameras, however – smart spaces in our cities now include sensors, cyber networks, and the monitoring of communications, exchanges and transactions for security operations and other purposes

However, with so much information being available from so many sources, how can we track and monitor it in real time to proactively prevent criminal acts? Much of the technology – cameras in particular – is used more as a deterrent, with criminal activity tracked down after the event to bring justice to those responsible.

Smart cities

A Europe-wide research initiative led by BU’s Professor Zoheir Sabeur is looking to change that, using AI to create truly smart cities, where CCTV – and other tools – can be used in a more proactive way to keep people safe.

“I’ve been working in the area of data science and artificial intelligence for the last 20 years,” says Zoheir. “When I came to BU in 2019, I brought my networks and contacts with me. Once of the first things I did was to pull together my understanding from a wide variety of areas, and I conceived the idea for S4AllCities pretty much at my desk.”

That idea was for smart cities to develop ‘digital twins’ where all the various sensing and monitoring equipment in a small area can be considered together, as a whole, with the support of AI. “The digital twin will enable us to use AI to look at what is going on physically in space, what is happening cyber-wise, what transactions are happening and so on,” explains Zoheir. “From there, we can start to understand if anything malicious is likely to happen.”

Stadium security

Zoheir gives football stadium security as a simple example. “I carried out some research in a previous role,” he says. “In a stadium that held more than 30,000 people, there are just a few guards looking at screens that relay feeds from dozens of cameras. How can anyone be sure they are looking in the right place at the right time?”

It’s into that space that solutions like S4AllCities arrive. “By using AI to assess crowd behaviour, it can draw the attention of the security operatives to one particular area where problems are likely to happen – whether it is crowd violence, emergency exits being blocked, or something else completely,” explains Zoheir. “S4All Cities is about bringing that physical analysis and combining it with communications, cyber attacks, transactions – everything we can measure, using AI to tell us where there may be a risk.”

Augmented intelligence

Zoheir describes the idea of artificial intelligence working alongside human interpretation as augmented intelligence. “The work is not about removing humans from the equation,” he says. “They are experts in their field and will always be the ones to make the final decisions. However, we can augment their support with AI  in order for them to focus only on the important tasks and avoid being distracted by unimportant factors.”

It sounds like science fiction, but Zoheir and his team’s work at BU and partners across Europe is making it science fact. The S4AllCities secured almost €10 million in funding. This has enabled his BU team to conduct its research in three locations – the cities of Trikala in Greece (one of Europe’s smartest spaces), Bilbao in Spain, and Plzeň in the Czech Republic.

“The aim is, I suppose, to revolutionise the way we protect our cities and their infrastructures,” says Zoheir. “The answer doesn’t lie just in more security forces, or  surveillance, but in smarter use of the security forces and surveillance we already have. The answer is the kind of augmented intelligence that S4AllCities is working towards.”