The use of facial recognition combined with large data sets and artificial intelligence (AI) will lead to personalised offerings whenever we walk into a bar, restaurant or shop. It will mean greater security and the detection of criminals or terrorist threats in vulnerable venues.
Facial analysis techniques offer great promise in the field of medicine. Many genetic diseases cause changes to the skull or face and facial analysis software is increasingly effective here – especially for remote diagnosis. Face2gene is an app from Boston start-up company FDNA which can use a face image together with other clinical data to assess the likelihood of a variety of diseases. It is used by doctors and researchers in over 100 countries.
Stanford researchers Kosinski and Wang developed a program which assesses sexuality (gay or straight) from face images. They fed in some 130,000 images from a dating website together with declared sexualities to build the algorithm. They claim that their program can identify whether a person is gay or straight from a single photo with 80% accuracy for men and 70% accuracy for women. The accuracy increases with more photos of the same individual. This performance far exceeds human capacity to judge this distinction. The opportunities for abuse of this capability are clear for example in the hands of a repressive government in a country where homosexuality is illegal.
And as the Economist pointed out in a recent editorial, employers could reinforce prejudices by using facial software to reject applications based on ethnicity, sexuality or assessed intelligence. If Facebook were to sell photos and data on its billions of users to governments or corporations then they would have enormous power to track, detect and anticipate the actions of people in all sorts of situations. This is a tremendous weapon against criminals but at the same time a huge threat to the privacy of individuals. The photographs of half the adult population of the USA are already held on databases which can be accessed by the FBI.
For these reasons many people are now calling for a ‘Digital Bill of Rights’ which would clearly establish and protect property rights in data, personal information, biometrics and face images. European regulators are proposing data-protection rules which include the provision that ‘faceprints’ belong to their owners and cannot be used without their consent. Users should be able to opt for anonymity or to move their entire ‘digital identity’ including photos, comments, likes, friends etc. when they move from one social media site to another. The site which was quit would have to wipe all of the leaver’s history.
Google has stated that it will not link faces to identities because of the risk or misuse. However, Amazon, Microsoft and Facebook all offer FRS in one form or another. Law enforcement agencies already make extensive use of facial recognition. We need to be planning appropriate levels of protection for individual privacy. It used to be said that your face is your fortune. If we are not careful our face could soon be someone else’s fortune and our misfortune.
Paul Sloane is the author of The Leader’s Guide to Lateral Thinking Skills and The Innovative Leader. He writes, talks and runs workshops on lateral thinking, creativity and the leadership of innovation. Find more information at destination-innovation.com.