What can your face say about you? Face recognition technology can pick up on things like your age, gender and maybe even your mood. Now, two researchers say it could even tell whether you鈥檙e a criminal.
They are claiming to have developed a system that, when shown a series of faces it has never encountered before, can pick out the ones belonging to convicted criminals.
But other researchers have criticised the results, and say the work raises ethical questions over what face recognition technology can and should be used to detect.
Advertisement
It鈥檚 clearly an 鈥渆motionally charged鈥� subject, says at McMaster University in Hamilton, Canada, who co-authored the study. He and his colleague Xi Zhang at Shanghai Jiao Tong University, China, had set out to disprove the idea that there could be a link between someone鈥檚 face and criminality 鈥� 鈥渟o we were very surprised by the result鈥�, says Wu.
The researchers exploited machine learning, asking face recognition software to guess whether a person in an ID-style picture was a criminal or not, and then feeding it the correct answer. It learned to tell the difference, eventually achieving an accuracy of up to 90 per cent, they say.
However, other face recognition experts question their methodology. One issue is that the criminal images came from a Chinese database of ID photos, whereas the non-criminal images were internet profile pictures belonging to Chinese citizens, meaning the system could have picked up on differences between the two sources rather than in people鈥檚 faces.
Wu and Zhang tried to counteract this by standardising the images, for example making them the same size and turning them greyscale. But from the Massachusetts Institute of Technology says that鈥檚 not enough. 鈥淭he fact that the data comes from two different places is a fundamental flaw. Any differences will be picked up,鈥� he says.
It鈥檚 not a problem to ask a controversial question, says , a deep learning researcher at Google, but the science has to be well founded. 鈥淚t is not ethical to make a bad science argument,鈥� he says.
at Falmouth University, UK, says that this kind of research risks turning machine learning into the 鈥減hrenology of the 21st century鈥�, like deducing a person鈥檚 traits from the bumps on their head. Seemingly impartial computer programs give an air of legitimacy to inaccurate or controversial interpretations. 鈥淪uddenly, the conclusions drawn by an algorithm have been cleaned up and made to look scientific,鈥� he says.
Same biases
In fact, these systems are not objective and are often subject to the same biases as humans. 鈥淸They] are tools that are forged by being hammered with our own beliefs and observations,鈥� says Cook.
That鈥檚 not to say computers can鈥檛 make accurate observations about a person鈥檚 face, sometimes even better than humans. Face recognition software can already easily pick up things like . Researchers at the University of Rochester, New York, even claim to have developed an algorithm that can with an accuracy of 75 per cent 鈥� significantly better than humans.
But even where the science is sound, ethical questions arise over how these algorithms should be applied to real-world situations. Detecting someone鈥檚 ethnicity, for example, could be used to better target services, but it could also be used to discriminate.
Last year, Microsoft that used machine learning to gauge someone鈥檚 age from their picture. It was intended as a 鈥渇un app鈥�, but then some UK newspapers used the system on images of refugees in an attempt to detect adults taking advantage of concessions given to children, forcing Microsoft to respond that it was never meant to offer a definitive assessment of age.
And when considering more complex or abstract characteristics than nose shape or age, it鈥檚 important to know the limits of what the technology can tell us. at Princeton University says you simply can鈥檛 glean someone鈥檚 general personality or behaviour from a snapshot of their face. It鈥檚 鈥渟uper easy鈥� to tell if a person is sleep-deprived based on paler skin and droopy eyes, he says 鈥� and this could even be used to prevent someone engaging in a task that requires alertness, such as operating dangerous machinery. 鈥淏ut if it is used to predict what the person is like in general, this is wrong.鈥�
Researchers do not always have control over how their work is used. Making findings public, as Wu and Zhang have done, means that anyone can scrutinise their validity, but it doesn鈥檛 have to be that way. 鈥淲hat would scare me more would be if a private company did this and sold it to a police department. There鈥檚 nothing to stop that from happening,鈥� says Frankle.
Earlier this year, that the majority of US police departments using face recognition do little to ensure that the software is accurate. As the technology becomes more widely used, so does the urgency of weighing up the ethics of its use.
Computer scientists are gaining increasing power over people鈥檚 lives, says Chollet, but they don鈥檛 have the ethical education to support that role. 鈥淭his is something we have to fix.
Reference:
