Two people in Ningbo, East China's Zhejiang province, were reported to have stolen a roommate's money from his bank account by unlocking his smartphone via facial recognition while he was asleep. Science writer Zhang Tiankan comments:

The local police have already solved the case. According to their analysis, the facial recognition of the victim's smartphone had a major flaw: it can function even when the owner's eyes are closed, which is uncommon for the facial recognition software of smartphones.

The case sends a warning about facial recognition technology, which has many flaws and deficiencies that are yet to be solved. The shape of a human face varies with the changes in light and observation angle. Wearing glasses, shaving, or a different hair cut, all these can result in the failure of human face recognition technology.

Besides, plastic surgery, losing or gaining weight, as well as growing older can all produce facial changes that make it difficult for the facial recognition software to recognize a person. If a criminal has the capability to make a mask that imitates someone's face, he or she can even cheat the facial recognition system by wearing it.

Human facial recognition technology is widely applied in sectors such as access to bank accounts and unlocking cars and smartphones. The police apply it when hunting for suspects too. But the technology is not mature and there are loopholes. It is necessary to further improve the facial recognition technology before putting it into wider use.

Even when facial recognition technology is fully mature and reliable, it is worth exploring whether it should be extended to all areas of people's lives, because, as with any technology, the question is whether people are to be the masters of it or slaves to it.