Facial recognition – whether it’s a passing fad, or the next big thing, one thing is sure: everyone has an opinion – good, bad, delighted or furious.
Your face has become a key. Right now, your face can be used to unlock your phone and approve mobile purchases but, in China, it links a person to their social credit score, which permits them to travel on public transport, or can ban the individual from travelling on trains and planes. Standing at the tip of Africa and looking up into the world, we can see the reactions to facial recognition and listen carefully to the debates. To our left there is a US city, San Francisco, that has banned facial recognition, and to the right, China, which embraces facial recognition without barriers.
But it’s not just commerce and security – there are also fun and sinister applications like using facial recognition data to create characters and avatars, or to create fake videos called deepfake, like the ones below.
What is facial recognition and how does it work?
Facial recognition is a method of identifying or verifying the identity of an individual using their face. Facial recognition is a category of biometric software that maps an individual’s facial features mathematically, and stores the data as a faceprint. The software uses deep learning algorithms to compare a live capture or digital image to the stored faceprint in order to verify an individual’s identity.
Safety vs privacy
The biggest debate is related to safety versus the ability to live anonymously. Safety is often raised as a reason to allow facial recognition to be used without regulation. Examples of this are many. Taylor Swift secretly used facial recognition to find stalkers in her concerts. But how accurate is facial recognition? The UK police have done several tests of facial recognition outside high-traffic areas such as high streets or football stadiums. These have led to some arrests, but only between one and three on each occasion, which – it is argued – an active police force could have achieved.
One of the greatest arguments for facial recognition is its potential use for finding missing people and combating human trafficking. Combining the ability to age missing people and continue to hunt for them globally using facial recognition has a lot of public support. Missing Children South Africa (missingchildren.org.za) reports that only 77% of missing children are found, and the use of global facial recognition could assist with lowering these statistics and with reuniting families.
The challenges of facial recognition
Failure rates are very high – recent tests by the UK police at the Notting Hill Carnival showed a 91% failure rate in identifying people (theguardian.com/uk-news/2018/may/15/uk-police-use-of-facial-recognition-technology-failure). Add to this a database filled with grainy photographs taken from surveillance cameras, and a list of offenders who have since been acquitted or wrongly accused, the concern is that often the police force is dealing with untangling wrongful arrests started with bad and/or old data, instead of keeping the public safe.
Questions to ask
Before investing in facial recognition, you should consider all the ramifications. There are so many questions, and every one leads not to answers but to more questions.
- How will data be collected?
- Will members of your community allow data to be gathered, not just of their faces, but of their children’s faces?
Who is on the database?
- Some companies share their data with other retailers and the police force, in order to use facial recognition to track offenders. Will all of the residents be on the database?
- Will they be linked to an identity, and how much information will that identity give (e.g. address, activities and other individuals they are linked to)? Linking a face to data means that a person’s exact movements and activities are recorded.
- How will children be handled? And visitors?
Where is the data stored?
- How will the data be stored to ensure that it is safe?
- Where will it be stored?
- What are the laws regarding the gathering, dissemination and use of the data? This is different in different countries. A company under great criticism for their use of data is Face Watch. Face Watch shares its data with private companies and the police. There are accusations that the data is also sold for monitoring a person’s shopping activities. This raises the question of who owns the data and how it can be used.
What mechanisms are in place to protect people?
- Corporates like Google are refusing to sell their facial recognition technology until they have the policies in place to ensure the protection of the public.
- In South Africa, there are interesting public conversations but the practice of giving specialists anonymity, as in this article in the Daily Maverick (dailymaverick.co.za/article/2018-09-28-joburgs-new-hi-tech-surveillance-cameras-a-threat-to-minorities-that-could-see-the-law-targeting-thousands-of-innocents/),raises deep concerns about accuracy and wrongful arrests.
In a country like ours with a history of pass books and surveillance, as well as a dangerous prison system, should a system with high levels of inaccuracy be utilised? And what is the cost to our society?
These are questions we must recognise and face up to…
About the author
Chantal Lailvaux is a co-founder of AIHO and Kin, as well as ambassador for the CityAI South Africa, and co-director for Pint of Science South Africa. Lailvaux’s passion is to bring simplicity to the delivery of STEAM-based complexity through research, projects and public discourse.