On December 12, 2019, Fortune Magazine reported that Kneron, an artificial intelligence company in San Diego, USA, has used high-definition 3D masks and photos to deceive China, including WeChat, Alipay, and Chinese train stations. As a result, some discussions on the security of face recognition have been initiated.
The Zero-ETL Approach: Enhancing Data Agility and Insight
What is the security risk of face recognition -C&T RF Antennas Inc
1. C&T RF Antennas Inc
www.ctrfantennas.com rfproducts1@ctrfantennas.com
Please Contact us for more information, thank you.
Jasmine Lu (86)17322110281
What is the security risk of face recognition?
Written By Calio Huang
On December 12, 2019, Fortune Magazine reported that Kneron, an artificial intelligence
company in San Diego, USA, has used high-definition 3D masks and photos to deceive China,
including WeChat, Alipay, and Chinese train stations. As a result, some discussions on the security
of face recognition have been initiated.
3D mask cracking face recognition is more a test behavior than a real risk
From the perspective of authentication technology, multiple dimensions are often constrained
together to verify one's identity at the same time. Our mobile payment actually has triple
protection. First, people don't give their phones to other people, they will be protected; second,
they usually set a power-on password. The protection provided by these power-on passwords is
very strong. Like Apple's privacy protection mechanism, even the FBI and CIA have no way. All
these levels have passed, and finally, it is WeChat and Alipay's face recognition. Behind this, there
are protection measures such as information about potential behaviors, common locations, and
time, and restrictions on payment amounts. Therefore, in this crack, the first few levels are
cooperated by the user, and then directly enter the test face recognition. Therefore, from the
perspective of security, it is more a test behavior than actual risk. Therefore, from the practical
level of practical life, consumers need not worry too much.
However, some thoughtful thoughts that have been expanded from this may wish to "concern" a
little.
Previously, a zoo in Hangzhou wanted to change the previous fingerprint admission certification
to face recognition certification. A professor at Zhejiang University of Science and Technology
who had already applied for an annual card brought the park to court because the reason was
that you did n’t ask for my opinion. By default, we all agreed to provide you with facial
information? The park was still quite confused at first, why do you think the fingerprint phone
number is not a big problem, but when it comes to "face recognition" everyone exclaims "my
privacy data has been stolen ".
Face recognition has higher data protection needs
The difference in public attitude is intuitive, because face recognition technology is the latest,
and people generally have more doubts about new things. However, from an objective point of
view, the risk of face recognition is far better than phone numbers and even higher than
2. C&T RF Antennas Inc
www.ctrfantennas.com rfproducts1@ctrfantennas.com
Please Contact us for more information, thank you.
Jasmine Lu (86)17322110281
fingerprints, irises, voiceprints, etc. that are also biological information.
The password is leaked. You only need to change the password, but the fingerprint and face
information cannot be changed. Once it is stolen and consciously stolen, the stolen person has no
other way except to change other security methods. Therefore, facial information is more
sensitive.
In addition, the appearance of deep fake technology allows face information to be used to
"replace the head" and port it to other videos. Deepfake is a combination of "deep machine
learning" and "fake photo". It is a successful application of deep learning models in the field of
image synthesis and replacement. Not long ago, the sensational "ZAO" face-changing APP storm
was based on deep fake technology. Therefore, for ordinary people, face recognition will bring a
completely new risk: Generally speaking, a password leak will only lose what is protected by the
password, but the spread of face information is likely to affect life.
Therefore, the European Union ’s data protection law, the “General Data Protection Regulation”
(“GDPR”) stipulates that facial images constitute “biometric data” under special types of personal
data, and are therefore subject to restrictions compared to ordinary personal data. Higher
protection requirements.
China National Internet Information Office, "Administrative Regulations for Network Audio and
Video Information Services" issued on November 29 also stipulates that from January 1 next year,
AI fake videos must not be released at will. There are four main provisions on AI fake audio and
video in the new regulations: "Perform security assessments in accordance with relevant national
regulations", "Identify in a prominent way" non-authentic audio information, shall not use AI fake
technology to release false news, and deploy AI fake audio and video Identification techniques
and sound rumor removal mechanisms.
The struggle between security and cracking security is the law of natural science
However, although face recognition is more sensitive, it also has its own advantages.
First, face recognition has great convenience.
In fact, to a large extent, face recognition is not a technology based on security considerations.
The passwords we usually use, such as payment passwords and bank card passwords, are hidden
in the user's mind, and there is no possibility of being stolen. If it is based on asymmetric RSA,
symmetric AES, and other encryption algorithms, as long as the password is slightly more
complicated, not simple words, mixed characters, letters, and numbers are basically unbreakable.
Therefore, in some very strict security scenarios, character passwords are used.
Of course, it may also be stolen when entering the password, but in contrast, the human face
cannot be hidden, and it is displayed in public places. Searchable on social networking sites.
Therefore, face recognition is not a completely security-based invention. However, on the other
hand, people's needs are diverse, and safety is not always number one, as many people think. For
example, for the convenience of travel, people will buy battery cars under budget constraints. If
this safety compromise is helpless, there are still people who spend more than 100,000 to buy
motorcycles instead of buying out of a hobby car. Therefore, safety is not the overriding
requirement, and convenience is also important. Therefore, in scenarios where both security and
convenience are required, face recognition is a very good solution. It still has great development
prospects in micropayments, company access control, home access control, time and attendance,
3. C&T RF Antennas Inc
www.ctrfantennas.com rfproducts1@ctrfantennas.com
Please Contact us for more information, thank you.
Jasmine Lu (86)17322110281
and smart home. of.
Second, face recognition has a wider range of objects. Face recognition, fingerprint recognition,
and other biological security measures, compared to password recognition and other schemes,
based on their own biological characteristics, also have some of their own unique advantages.
For example, if the password is to be impersonated, just tell others. But face recognition and
fingerprint recognition are even more difficult to counterfeit. The original owner of the face and
fingerprints cannot always be present at any time. Therefore, face recognition not only prevents
others from infringing on their rights but also protects the owner of the face to a certain extent.
For example, the face recognition of Didi Shunfeng and the courier is the platform to prevent
cheating.
Making face-changing videos and recognizing face-changing is like a cat-and-mouse game. One
foot tall and one foot tall, the contradiction between safety and cracking security has always
existed. The struggle between the two began with the invention of mechanical locks by humans.
From a deeper level, security and freedom are always a contradiction. When the security reaches
the extreme, freedom will inevitably be lost, and security cannot be maintained. Therefore,
counterfeiting of face recognition is itself a law of social development and even natural science.