The iPhone X that released in 2017 was enhanced with a new feature that became very popular. Now, we see this in almost all the new phones. Any guesses what this feature is?
Hint: It uses your face to unlock your phone.
That’s right! It is Face ID or Facial Recognition (FR)! What’s this? Software that identifies a person by scanning his or her face. The software captures your facial features, stores it in a database and later uses this to verify your identity.
How is it being used?
In medicine, FR is being tested to help to detect genetic diseases like DiGeorge syndrome. A person with this rare disease has slightly different facial features than someone who doesn’t. These differences are not very obvious. One study showed that the FR software was able to detect the disease with 96% accuracy by scanning a photograph!
There are also apps such as Seeing AI (Microsoft) to help people with lower vision to understand their surroundings better, see their documents more easily, read out text to them, and also help them to communicate better in social situations, like at a party. For example, it can identify people that the person with poor vision is talking to, and can help to analyze their emotions. For example, when the person he or she is talking to smiles, the phone buzzes! Isn’t that amazing?
Another way that FR’s ability to read emotions is being used is in some job interviews. In 2017, Unilever (the company that makes Dove soap) asked some job applicants for video resumes. These were passed through FR software to look for signs of lying, stress, eagerness, irritation, level of confidence, an “ideal” company face, and so on. Tiny facial and body movements can tell you a lot about a person. Sometimes, these tiny clues can be missed by humans but not by FR. The company used this information to hire some applicants.
Challenges: There are some serious challenges associated with the use of facial recognition, and people are questioning if it should be used in these kinds of applications. A lot depends on how well the software algorithms are coded, and on the people who have written those codes and trained the systems. We humans introduce our own biases into the software, which will then dictate the kind of accuracy it has and the kinds of mistakes that it makes.
Furthermore, the accuracy of the software will depend on how much data it is given – people come in all shapes and sizes, and so do our facial features. The software needs to be fed with large databases of features for people of different races and cultures in order to improve the accuracy rates.
As it relates to job interviews – there are certain things that people might explain in a face to face interview, that they can’t in a video interview with preset questions. These might have a bearing on whether or not they are given the position.
San Francisco Bans FR? On 14th May, 2019, the technology hub of San Francisco, CA, voted to ban FR from being used by government agencies like the police. FR is used by the police to search for criminals and to identify them in a crowd.
This sounds extremely useful so what’s the problem? FR doesn’t work properly for darker skinned people and for women almost 35% of the time. That’s a high error rate! What if you got arrested because you were mistaken for a criminal?
Question: Do you think that facial recognition software will be able to replace a human completely over time for certain applications? What might these applications be? Write to us in the ‘comments’ section below, or email us at: firstname.lastname@example.org with your thoughts.
Vaijayanti is a writer, a nature enthusiast and an amateur wildlife photographer. She hopes her virtual pen and lens can make the world a better place