Keeping smart cities safe with artificial intelligence

To address these issues, smart cities should implement security systems that work together to detect a threat, alert the right authorities, and address the issue before it turns into an event.

Smart security must be deployed in specific places, at specific times, with specific objectives, but without making the public feel uneasy. Strategic deployment can achieve a reduction in crime, especially when determining crime trends and/or crime displacement during large public events.

New artificial intelligence (AI) powered technologies can help in this by dramatically improving the effectiveness of today’s security systems, whilst blending in with their surroundings and not infringing on people’s privacy.

AI technologies are relatively straightforward to deploy across a city. They broadly fall into two categories:

1) AI with radar and magnetic technologies: Radar and magnetics are used with AI/machine learning to create multi-sensor, covert systems that can detect hidden physical threat objects, likes guns, knives, rifles, and metallic bomb material, on-body or in-bag.

2) AI with video recognition: Computer vision technology can be rolled out onto existing surveillance systems to detect visible threats such as knives, guns or rifles, or invisible threats, such as viruses. This can be deployed city-wide to identify and flag threats or suspicious behaviours immediately.

The use of AI removes human bias and error, which in turn addresses any civil liberties concerns. It is important to note that AI looks for the threat, not the individual, thus removing any suggestion of profiling. Highlighting that threat and passing the information to a monitoring station can then leave it up to the authorities to determine the response.

Whether it is used to tackle violence, such a fight or use of weapons, or to stop someone who could be risking the spread of Covid-19, AI systems can be easily programmed to spot the risks and, only when a threat is confirmed, reveal the details about the individual posing a risk to safety.

For example, in an airport, targeted magnetic sensors can be concealed in everyday objects, such as a decorative planter. As people walk past, the sensors scan the individuals and their bags for specified threats, such as weapons or explosives. Some sensors can even diagnose and track airborne trace explosives and chemical warfare agents. Furthermore, they could screen people for signs of Covid-19, ensuring travellers are healthy before boarding a plane.

The technology is already available and can enable an airport scanning model whereby passengers could pass through transportation hubs more freely. This could see a change to some of the airport security that we’re used to experiencing. Overt metal detectors and surveillance cameras that can feel intrusive and delay passengers could become a thing of the past.

As well as site-specific installations, city councils should look at the widespread deployment of a range of security technologies to make public areas as safe as possible. AI-driven technologies can be integrated into existing systems already installed in cities, such as sensor technologies, including radar, magnetic, video and chemical sensors.

These smart sensors work together to scan crowds of people and look for threats, without compromising a person’s privacy or civil liberties. Furthermore, smart sensors can be placed in strategic areas to help them blend in with their surroundings across a range of venues, including sports and music stadiums, schools and universities and large manufacturing plants.

Most people will pass these smart sensor technologies without even realising, but if they do pose a risk to safety, for example through the presence of a threat item or elevated body temperature, sensors can detect these and alert authorities with an image of the individual, so that appropriate response measures can be deployed to address the threat.

Real-time AI is now also being deployed to recognise behaviours that may put the wider public at risk. With smart AI, behavioural anomalies can be detected and even categorised.

Using video analytic software, traditional video cameras can be adapted to categorise a type of disturbance or to become remote video health screening tools. Example capabilities include identifying elevated body temperature, which could indicate a viral fever, or recognising if someone isn’t wearing a face mask or following social distancing guidelines.

Once a potential breach of health and safety compliance has been identified, the individual will be flagged to security personnel who can conduct a secondary screening with a temple thermometer or advise them on safety and social distancing measures.

There is still a mammoth amount of work to be done to keep citizens safe and AI must only be a tool to help law enforcement agencies; it cannot be there to replace human interaction. There are still those that are cautious around the use of AI, and civil liberty concerns and racial bias are issues that need to be resolved.

That said, AI used in conjunction with other threat detection technologies that are monitored as part of a safe city infrastructure has the potential to help keep the public safe in a non-intrusive way.