Recently, a video of Coco Lee being "resurrected" by AI has caused widespread discussion. Eight months after her death, an online blogger used digital technology to restore her smile and smile during her lifetime, with sounds and images generated by AI.
△ai generates video screenshots
In addition to Coco Lee, deceased stars such as Gao Yixiang and Qiao Renliang have been "resurrected" by AI. They wore different clothes, but when facing the camera, they almost always said hello first, and then expressed their thoughts about their fans and other heart-warming words. The publisher of
claims that this is a way of emotional comfort that uses high technology to provide emotional value to living people. However, in the video comment area, many netizens questioned: This kind of behavior is in the name of warmth, but in fact it is consumption. People who leave.
So, does using AI to “resurrect” the dead involve infringement? With the vigorous development of artificial intelligence, how can relevant laws and regulations be improved? Professor Zhang Linghan from the Institute of Data Law at China University of Political Science and Law explained ↓↓
Using AI to "resurrect" deceased celebrities, does it involve infringement?
Zhang Linghan said that the Civil Code has relevant provisions for the protection of the deceased's personality rights, including portrait rights, privacy rights and reputation rights. But now some new challenges have been raised regarding how AI infringes on the rights of the deceased, including the situation where the deceased is "resurrected" under AI technology. According to the traditional civil code, if fails to obtain the consent of the deceased's close relatives, it is an infringement. can regulate its legal liability by pursuing subsequent infringement liability.
In addition, in the "Internet Information Services Deep Synthesis Management Regulations" (hereinafter referred to as the "Regulations"), it is required that "deep synthesis service providers and technical supporters who provide facial, human voice and other biometric information editing functions should prompt the depth Users of synthetic services shall notify the individuals being edited in accordance with the law and obtain their separate consent.”
Zhang Linghan: Obviously, if fans want to "resurrection" a celebrity, it is difficult to obtain the informed consent of the deceased's family. There is no doubt that this is an infringement.
Regarding the governance of my country’s artificial intelligence field, Zhang Linghan said that my country’s existing artificial intelligence regulations are mandatory, and there are also systems in line with my country’s local characteristics. But even so, there are still legal deficiencies in the current governance of artificial intelligence. For example,
is mostly based on departmental regulations, so the efficiency level is relatively low. In addition, many regulatory authorities issue departmental regulations based on their own regulatory functions and lack an overall coordination mechanism and perspective. "The follow-up work to be done is to better coordinate the functions of these departments and better implement relevant governance."
How can standards and supervision keep up, or even get ahead?
In this regard, Zhang Linghan said that artificial intelligence technology is developing very fast, and making supervision and technology development as synchronized as possible sounds like an impossible task.
Last year, chatgpt developed rapidly, the number of users grew rapidly, and now many large artificial intelligence models have appeared. At present, the biggest difficulties faced by artificial intelligence governance are the rapid development of technology and the uncertainty of risks.
How to bring supervision to the forefront? A more reliable and feasible method is to always delineate the red line and bottom line of technological development, and always require that the development of technology cannot break through the bottom line of national security, social public interests and the protection of citizens' personal rights and interests.
In the face of AI technology, what laws need to be kept up with?
At present, artificial intelligence technology is developing rapidly and is creating more value for society. However, while artificial intelligence is changing our lives, there are also problems such as the spread of false information, infringement of personal information rights, and data security in its applications. For example, "AI face-changing" has spread from entertainment to fraud, and "AI singers" have been involved in infringement from entertainment.As new technologies develop, how can the legal system keep up with them?
In June 2023, the State Council's 2023 legislative work plan announced by the General Office of the State Council included the draft Artificial Intelligence Law for the first time in the scope of preparation for submission to the Standing Committee of the National People's Congress for review;
In September 2023, the legislative plan of the 14th National People's Congress Standing Committee was released , although artificial intelligence legislation is not explicitly listed in the legislative plan, the announced work deployment of the first category of legislative projects also mentioned the need to promote technological innovation and the healthy development of artificial intelligence.
Due to the rapid development of artificial intelligence, relevant departments are currently promulgating rules and regulations to deal with new problems that arise, and are constantly adjusting and regulating chaos in applications. What kind of constraints and reminders does the
ai "resurrection" technology impose on service providers?
Zhang Linghan said that my country has many regulatory requirements for producers. Among them, the "Regulations" state that if it can provide deep synthesis such as generating audio and video, it must first strictly manage data security, and at the same time, it must fulfill the obligations of to users, that is, When users upload sensitive personal information and edit it with faces and voices, they are prompted to obtain the consent of the relevant individuals.
At the same time, this is more difficult for platforms or service providers, because there are a large number of users generating different content, and if they are required to review them one by one, it will be difficult in terms of volume.
Even if it is difficult, providers of deep synthesis of Internet information services must still assume regulatory obligations. First of all, in terms of user agreements, relevant prompts must be done; secondly, when users upload sensitive personal information of celebrities, if they do not obtain authorization and consent, they need to perform relevant blocking and processing.
Source: CCTV News Official Account