Artificial Intelligence – The Metropolis is Watching You

Safer, smarter cities or Orwellian vision?

We’re still uneasy about the idea that Artificial Intelligence could be eavesdropping on our conversations through smart home devices, but what if AI is watching us, too? As unsettling as this might sound, AI computing company NVIDIA plans to set up over a billion AI cameras by 2020. Their new venture is called Metropolis, and can interpret a staggering 30 billion images per second. In future they plan to install these cameras in surveillance drones and robots. Following a $4 billion investment from SoftBank, NVIDIA are well placed to deliver a quality solution to current security camera issues. Integrating CCTV and machine learning has some clear implications for security and law enforcement, but which other industries could be disrupted by smart surveillance, and how will consumers react?

Intelligent watching

Machine learning techniques are already used as part of security protocol. Jaguar is working on a security system which recognises their owner’s walking gait, and FinTech startup Square allows users to pay for their coffee using facial recognition. AI has also been used in prisons to monitor inmate activity. Late last year, Altcourse Prison in Liverpool, UK, began trialling AI security cameras provided by Avigilon to stop contraband from getting into the prison. Although technology has undoubtedly helped to tighten security, the standard CCTV used outside of correctional facilities has some real set backs. Even if a criminal is captured by CCTV, the resolution of the image can be incredibly poor. And in the event of a crime or dispute, CCTV teams have to sift through hours of footage. This is obviously a tedious and time consuming job. However, a camera equipped with machine learning would complete the job in a matter of seconds, and images would presumably be of a higher quality. The cameras don’t just watch and record what’s going on – they can play an active role in their surrounding environment. In the event of road congestion, for example, the devices can help to redirect traffic. NVIDIA is just one company experimenting with AI surveillance. BriefCam, for instance, is a cloud based service already used in professional and domestic environments.

How will machine learning change security?

Applying AI to security might not be a new concept, but it has only been used in contained environments. NVIDIA’s Metropolis is different, because it has the capacity to track the movements of millions of everyday people. It sounds like the beginning of a Big Brother scenario, where drones and machines constantly spy on the population. Despite the negative connotations, smart surveillance cameras do have benefits. In a world policed by AI security cameras, people will be increasingly identified by their own physical traits, making it easier to prove ID and harder to fake it. As well as enhancing security protocols, Metropolis is in fact geared towards enabling safer, smarter cities – hence the name. Connected urban spaces need to be heavily coordinated to run smoothly, and machine learning equipped CCTV could offer a solution. Roadways, government property, public services and commercial buildings could all benefit from intelligent video analytics. In retail, marketers could use the cameras to gain valuable insights into customer preference and behaviour, leading to even greater personalisation in advertising. When it comes to crime, the availability of (almost) universal footage would significantly change legal procedures. Hard evidence would make it far easier for judges and juries to make decisions, but could have implications for professional lawyers by making their jobs somewhat obsolete. However, even if there’s a recording of a person blatantly committing a crime, lawyers will still be needed to negotiate sentences.

Whilst consumers might feel uncomfortable about being watched by AI, the ability to review events exactly as they happened is arguably a fair trade. If your car was broken into and the criminal was immediately identified by an AI CCTV camera, you would certainly see the merit in smart surveillance. In order for long awaited smart cities to function, operating systems will need to be acutely aware of changing environments. Even so, privacy is a touchy subject and companies like NVIDIA and Briefcam will need to be very, very careful about how they gather and analyse sensitive data.

Could your business make use of smart security and surveillance? Would you be comfortable knowing that AI was tracking your every move? Which other industries could be impacted by artificially intelligent CCTV? Comment below with your thoughts and experiences.