Recently, the Institute of Artificial Intelligence of Tsinghua University and its core incubator, Beijing Ruilai Intelligent Technology Co., Ltd., together with the Intellectual Media Research Center of Tsinghua University's New Transmission Institute, the National Industrial Information Security Development Research Center, and the Beijing Big Data Center, jointly released the "Ten Trends report on deep synthesis (2022)".
[The number of deeply synthesized videos released last year increased by more than 10 times compared with 2017]
Deep synthesis technology refers to the use of deep learning, virtual reality as the representative of the synthesis algorithm to produce text, images, audio, video, virtual scenes and other information technology. In 2017, a user named "Deepfakes" shared a pornographic video of a human face tampering with the face on the Us reddit website, bringing deep synthesis technology to the public.
The "Report" shows that in recent years, the number of deep synthetic content production and dissemination has increased rapidly. The number of deeply synthesized videos released in 2021 has increased by more than 10 times compared with 2017. In addition, the attention of deep synthesis content has also increased exponentially, taking the "likes/likes" interactive data of the video as an example, and the number of likes of newly released deep synthesis videos in 2021 has exceeded 300 million.
Since 2017, the number of papers in the field of deep synthesis has continued to grow at a rate of 30% per year. In addition to thesis research, open source projects in the field of deep synthesis also continue to climb at a rate of 30% per year.
The "Report" shows that the application scenarios of deep synthesis are constantly enriched, such as the restoration of historical photos, AI sign language anchors, virtual idols and other applications. More and more organizations are beginning to use deep synthesis technology to provide products and services for the public. Images and videos were most common in the early days of deep synthesis applications, and the number of them gradually decreased after regulatory norms were in place due to the uneven quality of products and the easy infringement of user privacy. In terms of audio, speech synthesis has become an important part of human-computer interaction, and is widely used in intelligent hardware, intelligent customer service, voice navigation, audiobooks, robots, voice assistants and other scenarios. In terms of text, deep synthesis has been more and more applied in poetry creation, chat Q&A, etc., and has shown creative efficiency and future potential.
The report predicts that with the gradual formation of technologies such as automatic data generation, whole-body synthesis, and 3D shaping, a new human survival scenario will be developed with deep synthesis technology as the cornerstone. The metacosm is a future human virtual digital space based on deep synthesis technology, which "completes the multiple replication and extension of real space and time, jumps out of the limitations of traditional physical space, and provides a new world where virtual people, natural people and robots are close to reality and transcend reality."
[The technical threshold is getting lower and lower, and illegal acts are not uncommon]
The "Report" believes that when deep synthesis technology penetrates into all areas of social life, the negative risks of deep synthesis content continue to increase and produce substantial harm. The technical threshold for the production of deeply synthesized content is getting lower and lower, and it is not uncommon to create false videos and false audio for false accusations, defamation, fraud, extortion and other illegal acts.
Human communication activities have also gradually entered the era of "deep post-truth" due to deep synthesis technology. According to the analysis of the "Report", first of all, "deep forgery" has profoundly affected the record of the truth, and the difficult identification of false content has affected the effectiveness of fact-checking. Secondly, if deep synthesis technology is used maliciously, it will use social media to cause false information to spread virally on the Internet in a short period of time. Third, in the release and tracking of information on daily events, deeply falsified information will also cause the continuous reversal of public opinion opinions, intensify the contradictions between different groups in society, and weaken the mainstream opinions on the social synergy. It is urgent to be vigilant that the malicious fake content of deep synthesis technology usually caters to the curiosity of the public and has a strong ability to shape consciousness.
The report also points out that the identification of deeply synthesized content is facing technical challenges. New forgery methods emerge in an endless stream, coupled with structural defects in detection algorithms based on deep neural networks, etc., identification technology is also facing "strong confrontation" and requires continuous updating and iterative optimization. This is similar to the "cat and mouse game", deep synthesis and identification will evolve themselves in the process of continuous learning to attack and defend, circumventing the confrontational techniques of the previous generation. At present, academia and industry at home and abroad have invested heavily in the research and development of identification and detection technology, and technology companies such as Ruilai Wisdom have launched detection products to support the detection of a variety of face-changing methods.
Countries around the world have issued laws and regulations to explore the legal management of deep synthesis. On January 28 this year, the State Internet Information Office issued the Provisions on the Administration of In-Depth Synthesis of Internet Information Services (Draft for Solicitation of Comments) and solicited public opinions, aiming to promote the rational and effective use of deep synthesis technology in accordance with the law and standardize the development of in-depth synthesis activities of Internet information services.
【Deep synthetic governance requires the joint efforts of law, ethics and technology】
How to standardize the application of deep synthesis technology and maximize the benefits and avoid harms, many industry experts and scholars expressed their views from the perspective of ethics and governance.
Zhang Cymbal, academician of the Chinese Academy of Sciences and founding dean of the Institute of Artificial Intelligence of Tsinghua University, said in the "Report" that the governance of artificial intelligence is a long-term task, not only to take the road of innovative development, adhere to the development of safer, reliable and controllable artificial intelligence; but also grasp the governance of artificial intelligence, from ethical norms, laws and regulations, industry consensus and other different levels to exert efforts.
Duan Weiwen, director of the Research Office of Philosophy of Science and Technology at the Institute of Philosophy of the Chinese Academy of Social Sciences, suggested that systematic and forward-looking interdisciplinary research should be carried out to carry out targeted governance and supervision of high-risk application scenarios that may occur in deep synthesis.
For the exploration of the governance path of deep synthetic content, Chen Changfeng, executive vice dean of the School of Journalism and Communication of Tsinghua University, said that efforts can be made from several aspects such as technology, ethics and legal system, and in terms of technology, through the participation of the government and social organizations, collaborative governance, and ethics, establish and advocate relevant deep synthesis principles.
Zeng Yi, a researcher at the Institute of Automation of the Chinese Academy of Sciences and an expert of unesco's Ad Hoc Expert Group on AI Ethics, advocated self-discipline and autonomy in industrial development, saying that before laws and regulations are fully mature and systematic, the industry itself should strengthen the awareness of "theory first" and jointly prevent abuse and strictly prohibit bad use as an industrial community.
Tian Tian, CEO of Beijing Ruilai Smart Technology Co., Ltd., also said that the essential problem of deep forgery is insufficient transparency, so it is particularly important to improve the public's awareness of deep synthesis technology, and only when the threshold is lowered to the time when all audiences can recognize, discuss and understand this issue in a common framework, deep synthesis technology can develop healthily and benignly.
The "Report" suggests that all parties should keep pace with the times to implement the new regulatory requirements, and under this premise, continuously pursue technological breakthroughs, continuously open up application scenarios of deep synthesis technology, create demonstration benchmarks, and form a driving effect on the artificial intelligence industry as a whole, so as to promote the sustainable and healthy development of new technologies.
![Cyberspace Administration of China: Deeply synthesized information content such as synthetic human face replacement must be conspicuously identified[fig]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)