Ten year challenges, facial recognition, and the future?
We’ve all heard the likely true rumors that Facebook and other social media have created the memeworthy ten-year challenge to perfect their facial recognition and age progression/regression technologies. But, have you considered why it matters to them?
Essentially it all boils down to data, and that boils down to $$$.
Facial recognition is big business is security. Currently Amazon’s Web Services’ Rekognition software is in use by several US states, and is being tested by the FBI. It has also been considered by ICE (Immigration Customs and Enforcement). All this is not even mentioning the enormous social engineering projects underway in China aimed at having every person in the country be facial recognized at all times in public, linked with a social credit score in the near future.
Facial recognition is also big business for big data. Facial expressions have always told us a lot about what a person feels and means. Dogs who have evolved to live with people have evolved to mimic and recognize our expressions as a means of evolutionary survival. So it should come as no surprise that our newest companion, AI, is learning to understand our facial expressions too.
Aside from age, which already tells advertisers mountains about you, facial recognition data science claims that it can tell much more about you from just a snapshot of your face. In German one such project has been able to identify over 200 distinct genetic disorders, just by shifting through snapshots of children’s faces. Another tool claimed to identify gays with a rate at 35% above chance (50/50), just by facial recognition.
The question is privacy and. increasingly. human rights. What if we find a tool that can identify potential criminals from their baby pictures with a rate of 90% accuracy? What if insurance companies shift through your profile pictures and deny you insurance because their software predicts that your face has a 75% chance of a certain genetic disorder? What if you loan is denied because a software doesn’t like your face? What if you are denied entry into college because of your face? What if less reputable governments use the software to discriminate?
If you think you have answers to these questions, you might be lying to yourself. Perhaps a facial recognition software can detect that?