Current position: Home > Media Trade > Text

Media Trade

"People's Daily (Overseas Edition)": (Zhang Xin) AI develops fast,tight "seat belt"

Release time: June 13, 2024 Editor: Wang Yan

(Source: "People's Daily (Overseas Version)" 2024-06-10)

The rapid development of artificial intelligence technology has caused traditional risks and emerging risks to superposition。Artificial intelligence governance has become an increasingly important topic。Recent,"People's Daily (Overseas Edition)" held a series of interviews,I interviewed a series of experts in this field on the issue of artificial intelligence governance。Professor Zhang Xin of our hospital was invited to interview and publish views。The following is the full text of the interview:

AI "Change Face" to implement fraud、AI forged pictures to spread rumors、AI "Resurrection" of the deceased caused controversy ... With the continuous development of artificial intelligence technology,Examples of AI in -depth forgery are common。Many netizens exclaim,I always said "there is a picture and the truth",Now we can’t even dare to believe in pictures and videos.。

What is the deepening of AI for falsification? How to prevent related risks?、Platform、Supervision、How to form a joint force,Establishing a format AI information security specification? We interviewed this。

w88 casino games

Hot events on the in -depth falsification of AI,It is attracting widespread attention all over the world。January this year,A popular singer in the United States was maliciously generated with AI for false photos,Quickly spread in social media,Causes trouble to the singer himself。February,A company in Hong Kong encountered "AI Change" fraud,Loss w88 online casinoof up to 200 million Hong Kong dollars。It is reported,An employee of this company is asked to transfer by the chief financial officer during the video conference。But,This "leader" and other employees at the meeting,Actually is a deep forged AI image。Information obtained by scammers through open channels,Synthetic image and sound of the chief financial officer,and make a false scene of multiple people participating in video conferences。

Infringement cases based on deep synthesis technology,The common method is to impersonate acquaintances to implement telecommunications fraud。Not long ago,Ms. Yang of Jurong, Jiangsu,After receiving multiple "daughters", after receiving the voice of multiple requirements to pay the registration fee,Transfer 35,000 yuan to the scammer account。The police handling police repeatedly reminded: "Be careful when encountering transfer,Seeing may not be real。”

"Deep fake technology uses AI deep learning function,Implement image、Sound、Video tampering、Forgery and automatic generation,Real effect with falseness。"Song Haitao, Dean of the Shanghai Artificial Intelligence Research Institute, brought three moves to the public in depth: First, master the simple technique of identifying AI" face change ",For example, asking the other party to wave in front of the face during a video dialogue,See if image disturbance occurs, etc.; The second is to learn to use the tools and software of depth for falsification; the third is to maintain reasonable doubt。"Keep caution and vigilance,is the first line of defense for the public to deal with AI fraud。"Song Haitao said。

"Technical Governance Technology"

How to use the powerful capabilities of generating formula AI to establish information security specifications,Turn the "stumbling block" that restricts the development of the industry into a "cockpit stone"? The industry has been exploring the "technical governance technology" scheme。

Ruilai Wisdom is an enterprise hatched by the Tsinghua University Artificial Intelligence Research Institute,Specializing in the field of artificial intelligence security。Introduction to Tian Tian, ​​President of Ruilai Smart,The generated artificial intelligence content detection platform developed by its company,Supporting w88 casinoa variety of synthetic types of pictures、Video、Audio、The authenticity of the text,Application scenarios include cracking down on cyber fraud and reputation violations、Detecting network content synthesis、Testing audio and video physical evidence authenticity, etc.。

"Use AI technology to control AI crime,Itself is also a continuous game。"Tian Tian Introduction,"Red Team Test" is an important means of generating AI governance at present,It aims to use analog attacker behavior,A comprehensive network attack on the target system,Targeted discovery、Repair potential system loopholes,Make the model before opening to the public,Fully accept security technical test。Video synthesis AI platform sora,Before listing, several experts engaged in information security vulnerabilities were invited to act as the Red Team for confrontation test,Find a considerable amount of security vulnerabilities。

In addition,The industry is also promoting the implementation of the AI ​​generating content identification system。Zhang Linghan, a professor at the Professor of Data Law of China University of Political Science and Law,The National Information Security Standardization Technical Committee gives the content identification method in the "Conducting Guide of Network Security Standard Practice -Generating Artificial Intelligence Service Content Identification Methods": Display watermark logo by adding translucent text to the interactive interface,Or the hidden watermark logo that can be directly extracted from the content through the technical means,The prompt content is generated by artificial intelligence。"The identification system can improve AI information content governance capabilities,Reduce false information generation,Prevent false information from polluting the next generation of training data,Create a good network information ecological environment。"Zhang Linghan said。

Founder of the International Testing Committee、Researcher Zhan Jianfeng, Researcher at the Chinese Academy of Sciences,AI in -depth falsification into the monitoring mechanism,False behaviors with a large negative impact,Quick reaction as soon as possible。The mass reporting mechanism for the establishment of in -depth falsification w88 casino gamesof harmful content,Improve the public's judgment、Differential Force。

"Institutional Guidance Technology to Good"

Genesis AI technology is a double -edged sword,How to release the vitality of innovation,Effective prevention of information security risks? Many experts say,Establish and improve related governance specifications as soon as possible。"The governance in this field cannot be completely hoped to be in corporate autonomy,More legal hard supervision。"Zhang Xin, director of the Digital Economy and Legal Innovation Research Center of Foreign Economic and Trade University, said,"Defining the basic bottom line for AI development through legislation,Clear compliance obligations,It can prevent risk from the source,Avoid the passive situation of ‘first pollution and then governance’。”

The Interim Measures for the Management of Genesis Artificial Intelligence Services issued by the State Cyber ​​Information Office and other departments last year marked that the generation AI has a specialized administrative regulations。The "In -depth Synthetic Management Regulations of Internet Information Services" implemented in January 2023 is clearly proposed,"No organization or individual may use in -depth synthetic services、Copy、Published、Communication Law、Information prohibited by administrative regulations "" may cause public confusion or misunderstanding,The reasonable position of the content content of the information content of the generation or editor、Regional significant identification ",etc.。

At the legal level,China has introduced laws such,But there is still room for further refinement of judicial governance in the AI ​​field。"Existing laws and regulations on data ownership、The scope of protection has no specific regulations。In terms of the protection of the result of artificial intelligence,,The principles of "Civil Code" still have certain difficulties in specific applications。"Suggestion Yan Jun, the person in charge of the comprehensive trial of the Beijing Internet Court,To further speed up the legal supply in artificial intelligence,Promote the establishment of artificial intelligence specifications at the national level。Difficult legal issues on existing cases,To discuss and practice w88 casino games loginthrough academic discussions,Gradually form a consensus,Summary and establish a rule of law,Promote the legislative process of the "Draft Artificial Intelligence Law"。

At the ethical level,AI "Resurrection" and other behaviors caused some users' resentment and fear。For this,The Ministry of Science and Technology and other 10 departments issued the "Science and Technology Ethics Review Measures (Trial)",list a list of science and technology activities that require ethical review and review,Among them, algorithm models with social mobilization and social consciousness guidance、Application and System R & D, etc.。

"We must correctly look at the possibility brought by the new format,But it cannot be partially clutch、Reasonable track。Whether it is technical developer、Users are still regulators,It is necessary to use a similar ethical perspective to examine technological development。While pursuing scientific and technological progress,Make sure that AI technology follows the concept of people -oriented and technology to good development。"Zhang Linghan said。

w88 casino games

Link to the original text:https://m.huanqiu.com/article/4I9FQdFc7af

w88 live casino

  • w88 casino games login
  • w88 online casino

Submitted mailbox: news@uibe.edu.cnReader's feedback: xcb@uibe.edu.cnAll rights reserved CopyRight of the Propaganda Department of the Party Committee of the University of Economic and Trade
School filing number: Foreign Trade Network 31418006