Artificial intelligence has long been watching our online and data selves. Now, it’s watching us in the physical world too. Facial recognition technology has given eyes to AI, allowing it to ‘see’ and analyse us in the flesh.
This ability has met obvious reservations. Indeed, concerns abound that we are heading for an Orwellian dystopia. Big Brother could soon exist in the form of AI — monitoring our actions both on and offline.
But facial recognition tech is also beginning to prove useful for businesses. It holds the potential for better safety, security and personalisation. So, what does the future hold for the fledgling AI branch? Howard Williams from Parker Software investigates.
The state of facial recognition
Facial recognition is one of the biggest buzzwords of the decade. Facial recognition AI uses machine learning to allow computers to recognise and verify human faces. As AI and image recognition advances, so too does the accuracy of facial recognition.
Some claim that facial recognition technology is already more accurate than most might expect. In fact, it’s achieved scores of 98.52 per cent accuracy in one study. However, this is dependent on specific criteria, such as clear images of a single person.
In crowds, accuracy drops significantly. There are also reports of issues with AI identifying people of colour. Not to mention how easy it is to fool. Indeed, it’s much easier to find reports of facial recognition AI with 98 per cent inaccurate results. And this is a cause for some concern.
Heading for a dystopia
Beyond the natural change aversion that tends to accompany new technology, facial recognition feeds more serious concerns. This is down to the fact that its use is already growing in law enforcement and surveillance. Despite, that is, its accuracy issues.
The concerns and fears around the current state of facial recognition AI boil down to ethical issues. Namely, the apparent loss of privacy brought by use of the tech. (Particularly in the case of surveillance.)
The issue is, for facial recognition AI to work, it needs images of faces to compare and learn from. In some cases, this might be as simple as temporarily storing images of your face as you enter certain places. For example, if the AI needs to confirm two images are of the same person.
Already, then, you could have cameras and computers ‘watching’ your movements throughout the day. However, for it to then identify you, it needs access to stored data about you. So, there’s some personal information required. The result is a scary public introduction to facial recognition AI. It’s generating the fear that Big Brother AI will soon be watching our every move.
Other ethical concerns
Privacy isn’t the only concern that has people worried about facial recognition AI. There are also issues surrounding the training of the system, and the potential outcomes of relying too much on the technology.
For example, there’s an issue with where and how companies collect the training data for the AI. The worry is that many data sets were collected without explicit permission from the individuals involved. Some cases have allegedly seen face data scraped from photo apps, for example.
The issue is, many official face data sets don’t hold enough data or diversity to teach an AI system. This means without more faces to feed to the AI, you get a poor tool with plenty of bias.
Another concern surrounds the issues of accuracy. Namely, the repercussions of false positives if law enforcement relies too heavily on inaccurate AI. What happens to an innocent person facing accusations due to an AI misidentification?
The potential of facial recognition
On the flipside of the coin, facial recognition technology could take a different route. Stepping away from the dark side of facial recognition AI use, the future stands to be much brighter. Already, businesses are starting to see the value in facial recognition.
In fact, it’s a technology that could prove beneficial no matter the industry you’re in or what your business offers. It supports security, it’s useful in healthcare, and boosts personalisation in retail. As a new technology, the scope of its ability is still growing. Indeed, facial recognition AI could become an extremely versatile tool.
Safety, security and personalisation
Facial recognition AI could boost the safety of your business. For instance, it provides a potential way to combat theft. The computer system could scan faces as they enter the store, and check against a database of known offenders.
Or, the technology could provide a new form of authentication. Keys to your buildings are no longer lost if people unlock the doors with their face, for example. Here, facial recognition AI could recognise the face of someone trying to unlock a building or device. Then, it could check the image against a database of your employees. This would ensure only those authorised can gain entry.
But facial recognition doesn’t just need to revolve around security. It can also boost your personalisation efforts. You could have a system that recognises customers by face as they walk into your shop. Then, it can pull their data from your CRM to help your employees tailor service to them.
Facial recognition AI acceptance
So, what needs to happen for people to accept this form of AI use?
- Address privacy concerns
Be transparent about the use of facial recognition AI and why you need it. Keep clear how it benefits the consumer. Make sure that you have permission to collect their facial data and make it clear how and when you will delete it.
- Remember it’s a tool
It’s important to remember that (at least for the foreseeable future) facial recognition AI is not infallible. Plus, as with any AI, the output you receive is best taken alongside human understanding. In other words, use it as a tool to inform decisions, not a brain to make them.
Approaching a crossroads
We stand now at a crossroads. In one direction, an Orwellian dystopia with no privacy and total surveillance. In the other direction, a reality of increased personalisation and embraced individuality.
It’s true that the technology could lead to a real Big Brother. But it can also revolutionise our businesses. It could improve security and add another layer of personalisation to customer service.
Facial recognition AI is a tool. It's no different from other functions of artificial intelligence, or automation, or any other software. And, as with any tool, how we use it is up to us.
Howard Williams, customer experience, Parker Software
Image Credit: Sergey Nivens / Shutterstock