When the face recognition station is on the air, the controversy has never stopped.
Nowadays, people pay for face brushing, check face brushing, check face brushing in hotels… Even though face recognition technology has been applied in many aspects, there is still a technical dilemma. At the same time, with the rapid development, the voice of reflection gradually increased: face recognition, where is the boundary? What “rules of the game” should we abide by in balancing privacy, security and convenience?
Beijing News reporter talked to a number of experts about this. Among them, Wu Zhuoqun, President of Anheng Information Security Research Institute, said that new technologies always have security problems. Face recognition itself provides convenience for life, and its biggest risk is information leakage. Wu Shenkuo, senior adviser on cybersecurity and cybercrime of the United Nations, told the Beijing news that once face recognition technology is popularized, it can locate someone who is not aware of it. However, the security problem is not the technology itself, but the application.
What are the advantages and disadvantages of focus 1 face recognition technology?
Wu Zhuoqun: face recognition belongs to a kind of biological recognition. Police case solving and face changing software are all application cases of face recognition technology. The biggest advantage of this technology is convenience, so it is popularized quickly.
However, there are also some phenomena to attack and crack the face recognition mechanism by fighting against the network. As a kind of biometrics, face is unique and has great risk of information leakage. Therefore, it is necessary to ensure the security of information flow from the institutional level and restrict the collection of face data from the legal and institutional level to prevent information leakage.
In a word, face recognition itself is convenient for our life, which can avoid the trouble of password input, and its biggest risk is information leakage, so we need to constantly improve the algorithm, improve the law and system.
Majie: at present, face recognition technology has been widely used in intelligent door lock, mobile payment, mobile phone unlocking, etc., which improves the safety factor in the aspects of secret theft, fraud, false use, replacement, etc. In the company, shopping malls, airports, schools and other scenes, face recognition technology can improve the management efficiency and enhance the interactive experience.
In addition, in the criminal investigation and public sphere, face recognition technology can help to capture suspect who absconded for many years and help family members who have been separated for many years, so that the efficiency of law enforcement and social public security can be greatly improved.
But correspondingly, as a kind of biometrics, face recognition technology is easy to be cheated (cracked).
Focus 2: with the popularization of “face brushing”, will the hidden danger be greater?
Ding Xiaodong: I’m against one thing in face recognition. I think face recognition technology is sensitive and can’t be touched. In many cases, human faces are public data. We can be seen on the street every day. We can’t say that human faces can’t be shown to others. In fact, our face information has been circulating and flowing with each other for a long time, so I don’t think we can take a prohibited attitude to this technology just because of Zao incident, but we should take precautions against risks in specific scenes, which may be a more important and more necessary attitude.
As far as the dispute is concerned, the point is not whether the face change itself will have problems, but whether the public can form a reliable identification of this technology. For example, when we read comics, we know it’s fake. If the use of some technologies makes it difficult for the public or ordinary people to identify them in a short period of time, there may be risks, such as changing the face of the news anchor, and then creating fake news.
Face recognition may cause great risks in some areas, including the news field. Through deep forgery technology, forged news pictures can be spread.
Wu Shenkuo: technology is neutral, but people who use technology have a position. The security problem of face recognition technology is not the technology itself, but the application problem. The danger lies in the use of this technology to achieve the goal should not be achieved, the implementation of the behavior should not be implemented.
Once face recognition technology is popularized, it can locate someone who is unknown. The technology used in the investigation of crime is to help public security; When it is used to track others, it may violate privacy and interfere with others’ freedom; If it is applied to impersonate others, it is committing a crime. Therefore, whether face recognition technology is harmful or not depends on how to use it.
Can’t the security of focus 3 face recognition technology be guaranteed?
Zhang Baichuan: there will be risks when face data is widely used, and there is no way to avoid network black production. When face recognition first appeared, taking a picture and printing it out may crack it. Now you may notice that some software will ask you to blink and nod your head, but it may also be cracked. But I also want to say that fingerprint identification is also risky. There have been cases of fingerprint data leakage.
Maggie: the so-called “cracking biometrics” is a process of “deceiving sensors”. Nowadays, more and more intelligent devices adopt biometric technology, but we find that most of these biometric technologies have the security risk of “sensors are easy to be cheated”. Baidu security research shows that although biometrics has become more and more mainstream, from the perspective of security, it is not foolproof.
But you don’t have to worry too much. We have basically found the corresponding solutions for the loopholes found before. We invest a lot of energy in the research of biometric security. On the one hand, we call on equipment manufacturers to consider product experience and security at the same time. On the other hand, we discuss with manufacturers how to solve these problems and promote the progress of biometric technology.
Chen Litong: face recognition technology is originally used to verify “whether you are really what you say you are”. In the case of remote control transaction or identity confirmation, confirm that an action is implemented by myself. It is the most effective technical measure to solve the identity security problems such as electronic payment, network transaction and network application for public services. However, once the face information, which is easy to obtain “biological code”, is lost or replaced at will, it will be ridiculous to use face recognition technology to verify the real identity.
Focus 4 how can the public protect privacy?
Wu Zhuoqun: because the way of face recognition is to collect face data first, and then to match and verify, face recognition institutions have actually collected a lot of facial features, original pictures and photos, which will be kept on their servers, so it is difficult for ordinary people to avoid the risk of face recognition. What the public can pay attention to is not to provide photos to very small face recognition institutions.
At present, the cases of face recognition being exposed and abused are relatively few, but face changing and other attack methods have been more common. For face recognition, in fact, every company has different algorithms. Simultaneous interpreting faces from different sensors through different algorithms is different. Therefore, the same photo can be recognized successfully in one home, but not necessarily in another.