Xinhua held an invitation testing press conference for the "AIGC Safe" platform for generative artificial intelligence content security and model security detection

Tech 2023-07-13 13:29:41 Source: Network

The "AIGC Safe", jointly developed by Xinhuanet and the Institute of Computing Technology of the Chinese Academy of Sciences (hereinafter referred to as "the Institute of Computing Technology of the Chinese Academy of Sciences") and other industry institutions, invited the test conference to be held in Beijing a few days ago. This platform aims to collaborate with industry authoritative ecological partners to jointly strengthen the supervision of the AIGC industry, and play an important role in the safety governance and healthy development of the AIGC industry

The "AIGC Safe", jointly developed by Xinhuanet and the Institute of Computing Technology of the Chinese Academy of Sciences (hereinafter referred to as "the Institute of Computing Technology of the Chinese Academy of Sciences") and other industry institutions, invited the test conference to be held in Beijing a few days ago. This platform aims to collaborate with industry authoritative ecological partners to jointly strengthen the supervision of the AIGC industry, and play an important role in the safety governance and healthy development of the AIGC industry.

Invite the testing press conference site. Photographed by Liu Xia on Xinhua News Agency

Against the backdrop of accelerating the integration of the digital and physical worlds, generative artificial intelligence is reshaping or even disrupting the production and consumption patterns of digital content, increasingly influencing changes in various industries and becoming an indispensable support force for entering a new era of digital civilization in the future. Since this year, the rapid integration and development of big model technology and AIGC have provided a solution to meet the personalized production of massive content with low Marginal cost and high work efficiency from the technical level, promoting the content industry to develop in a more creative direction. However, the threats and challenges to national security, economic development, and social stability caused by its abuse have also emerged, and the regulation of generative artificial intelligence services has become a major issue in global governance.

Zhang Ruining, Vice President of Xinhua Network, delivered a speech. Photographed by Liu Xia on Xinhua News Agency

Zhang Ruining, vice president of Xinhuanet, said in his speech that as a central key news website hosted by Xinhua News Agency and an Internet listed enterprise with strong innovation strength, Xinhuanet continued to promote the deep integration of technology and media, actively served the national digital economy strategy, and fulfilled its mission. The China Internet Joint Anti Rumor Platform, hosted by the Central Cyberspace Information Office's Illegal and Adverse Information Reporting Center and hosted by Xinhua, aims to "publish authoritative anti rumor information, improve netizens' media literacy, and create a clear online space". It has initially formed a work system of "linkage discovery, linkage clarification, linkage disposal, and linkage anti rumor", and has become the main source, channel, and battlefield of authoritative online anti rumor information. The AIGC Safe released during this test is an important exploration by Xinhua Network in AIGC security governance. This platform is conducive to strengthening the detection and prevention of AIGC multimodal deep fake content, using AI to protect AI, and safeguarding the better development of the industry.

Wang Fan, a member of the Party Committee of the Institute of Computing, Chinese Academy of Sciences, delivered a speech. Photographed by Liu Xia on Xinhua News Agency

Wang Fan, a member of the Party Committee of the Institute of Computing, Chinese Academy of Sciences, stated that the AIGC-Safe test released by Xinhua Network is an attempt to build a scientific research model of "production, learning, research, and application", and is an innovative move to test technology through industry implementation. The Institute of Computing Technology of the Chinese Academy of Sciences will fully provide core technical support for the platform and contribute to the healthy development of the AIGC industry.

Yuan Junpeng, Chairman of the Scientific Research Integrity and Responsible Innovation Professional Committee of the Chinese Academy of Sciences and Science and Technology Policy Research Association, and Deputy Director of the Scientific Research Integrity Research Center of the Documentation and Information Center of the Chinese Academy of Sciences, delivered a speech. Photographed by Liu Xia on Xinhua News Agency

Yuan Junpeng, chairman of the Scientific Research Integrity and Responsible Innovation Professional Committee of the Chinese Academy of Sciences and Science and Technology Policy Research Association and deputy director of the Scientific Research Integrity Research Center of the Documentation and Information Center of the Chinese Academy of Sciences, said that while AIGC improved efficiency, it also posed challenges to the scientific research ecology. He looked forward to working with all sectors to jointly promote the development of scientific research and create an honest, fair and transparent scientific research environment with accurate and reliable AIGC detection products.

Cao Juan, Chief Technical Expert of AIGC-Safe Platform, tested and released AIGC-Safe. Photographed by Liu Xia on Xinhua News Agency

Cao Juan, the chief technical expert of AIGC-Safe platform, conducted on-site testing and release. The platform products are based on the digital asset and data element management technology base of Guodian Chain (Guodian Chain), and are built with the technical accumulation of the Institute of Computing, Chinese Academy of Sciences. It has formed two core capabilities of AIGC deep fake content detection and model detection, and can open up and empower various AIGC detection business scenarios. Model security ensures the entire process from training to inference from three aspects: training data security, model attack prevention, and model input security; Content security covers the detection of text, images, and audio and video, ensuring the authenticity and compliance of content, and achieving dual security protection.

This AIGC Safe platform test release focuses on the online content security detection function, supported by the leading detection model and deeply forged content Big data. Text detection supports AI generated authentication; Image and video detection can cover deep synthesis forgery such as facial generation, facial editing, facial replacement, and expression transfer, as well as AI generation and PS tamper detection; Audio forgery detection supports TTS and VC audio synthesis detection, covering mainstream audio synthesis algorithms. The content security function of AIGC Safe platform is fast, accurate and easy to deploy, which effectively reduces the cost of industrial security technology investment and improves the efficiency of security protection. It can be widely used in a variety of detection scenarios, such as fake news, AI facelift fraud, live attack, copyright content protection and academic integrity, and can escort AIGC security governance in media, education, finance, public security and other fields.

At the press conference, an invitation testing session was set up, and AIGC-Safe conducted online testing on multimodal AI generated or forged materials such as text, images, audio, and video provided by attendees. The testing scope covered mainstream AI generation algorithms, with excellent performance in accuracy, response time, and other aspects, which was unanimously recognized by the on-site guests.

Round table dialogue session. Photographed by Liu Xia on Xinhua News Agency

The press conference also held a round table dialogue with the theme of "Protecting AI with AI". Cao Juan, Yuan Junpeng and Jing Ke, the director of the operation center of the China Internet joint rumor dispelling platform, Chen Jing, the deputy manager of the middle platform architecture department of the information center of China Southern Airlines, Wang Hui, the ecological director of Tencent Smart Movie, and Lei Tao, the director of information security products of Orange Shield Information Technology Co., Ltd. (Alibaba Holdings), shared wonderful views. Industry experts and representatives from institutions such as Baidu, 360, Weibo, Zhihu, CNKI, Guodian Chain (Guodian Chain) super nodes Yirong Technology, and Zhongke Ruijian attended the meeting.

It is reported that in the face of multiple challenges such as rapid technological iteration and the continuous emergence of new business scenarios in the AIGC field, AIGC-Safe will collaborate with industry management agencies, large model research and development institutions, news media units, digital scene application platforms, and other partners to initiate the construction of the "Generative Artificial Intelligence Content Standardization Governance Development Ecological Matrix", escort digital technology innovation and application, and create a "new paradigm" for digital content security governance.

Disclaimer: The content of this article is sourced from the internet. The copyright of the text, images, and other materials belongs to the original author. The platform reprints the materials for the purpose of conveying more information. The content of the article is for reference and learning only, and should not be used for commercial purposes. If it infringes on your legitimate rights and interests, please contact us promptly and we will handle it as soon as possible! We respect copyright and are committed to protecting it. Thank you for sharing.(Email:[email protected])

Mobile advertising space rental

Tag: for security Xinhua held an invitation testing press conference