With Facial Recognition, Does AI Take Away All Our Privacy?


Advancements in AI and machine learning are helping us make smarter decisions, in turn making our business processes more efficient. They are also helping us automate our day to day tasks, making our lives easier. One such task is facial recognition. We use it to unlock our phones, send filtered pictures to one another and even see how we would look when we’re older. But digging deeper, we need to ask ourselves, does all this take away from what little privacy we have left?

A virtual guardian for your store

Take the AI Guardian for example. This is a security camera by NTT East and Earth Eyes Corp designed to help store owners in Japan identify potential shoplifters.  Using open source technology, developed by Carnegie Mellon University, the AI Guardian can scan a live video stream and guess the pose of bodies that it can see.

These poses are then matched against a database of poses that are predefined as suspicious. If a customer exhibits this type of behavior, then the shopkeeper is notified via AI Guardian’s smartphone app.

According to NTT East and Earth Eyes Corp, the AI Guardian has reduced shoplifting in stores by around 40%. The AI Guardian is expected to go on sale by the end of July 2018, priced at around $2,150 and a monthly subscription fee of $40 for cloud support. This could be used to update databases of postures so that the system can improve identification of suspicious activity. NTT hopes to implement the AI Guardian to 10,000 stores within the next 3 years.

That being said, there are limitations to the system. The primary being the identification algorithm itself. For example, even if AI Guardian can look at a person and match the locations of their body parts, using that to identify suspicious behavior is challenging.

As such NTT East did admit that the AI Guardian misidentified a few customers. These customers were indecisive in their items and picked up an item, put it back and then picked it up again. AI Guardian flagged this as suspicious behavior. Another example was that of salespersons restocking shelves as well. Would you be okay if you had a guard following you because you were “acting suspiciously” based on what a camera told you? Would you go back to that store again?

AI is getting smarter at identifying skin tones

There are also possibilities that the AI might be biased towards certain groups of people or that it might even discriminate against them. For example, Microsoft reported that their facial recognition tools are getting better at identifying people with darker skin tones. Rates of errors have been reduced by as much as 20 times for men and women with darker skin and by nine times for all women.

Microsoft has reported that their facial recognition tools are getting better at identifying people with darker skin tones (Image Credits: VentureBeat)

In case you didn’t know, In February 2018, a report from MIT’s Media Lab tested the facial recognition systems from Microsoft, IBM, and China’s Megvii. The results showed that up to 35% of darker-skinned women had their gender misidentified by the systems. Facebook and Google have also taken strides in face tracking. In fact, have you noticed that the moment you upload a picture on Facebook, it suggests people to tag in it? Well, that’s Facebook’s Ai at work.

Smarter law enforcement thanks to AI

Fixing Surveillance cameras across a country takes tremendous time and effort. It also takes manpower to monitor all the cameras. So what about a more portable solution? Well, that’s where China is excelling. The country operates the world’s largest surveillance state. That’s 170 million CCTV cameras at work. To make things easier for law enforcement, smart eyewear has been developed for police officers.

Resembling Google Glass, the smart specs connect to China’s state database to identify potential criminals by means of facial recognition. The specs are able to identify people in a crowd by taking their picture and matching it to the database. Officers are then presented with the person’s name, and address as well.

Resembling Google Glass, the smart specs connect to China’s state database to identify potential criminals by means of facial recognition (Image Credits: The Verge)

These advancements are not without their corresponding fears as well. For example, with the AI Guardian, will we face a world where everything we do will be monitored by a camera? Will we not have free will or privacy anymore?

Consider the fact that China is an authoritarian state. Say something or do something wrong and you will face the full force of the law. If you are a criminal, you can’t hide because all your movements are tracked by smart cameras. But what if you’re innocent and were in the wrong place at the wrong time? What would you do if you were arrested for a crime you didn’t commit? Well, you’ll have a lot of explaining to do.

With Sri Lanka’s ties to China, especially with regard to the developing Port City, it might just be a matter of time before these technologies are implemented in Sri Lanka as well. If so, then what? Will we finally have a crime-free Sri Lanka? Or will people find ways to beat the system, as they’ve been doing before?


Please enter your comment!
Please enter your name here