China's new regulation aims to label AI-generated content with explicit and implicit watermarks.,The policy holds social media platforms accountable for identifying and labeling AI content.,China's proactive stance on AI regulation could influence global standards.
The Race to Regulate AI
Artificial Intelligence (AI) is changing the world rapidly, and governments are racing to keep up. China, known for its swift tech advancements, is now taking a bold step to regulate AI-generated content. On September 14, China’s Cyberspace Administration drafted a new regulation to ensure people know whether content is real or AI-generated. This move comes as generative AI tools become increasingly sophisticated, making it harder to tell what’s real and what’s not.
What Are AI Watermarks?
AI watermarks are labels that indicate content is AI-generated. These can be explicit, like visible watermarks on images or sounds of Morse code before audio clips. They can also be implicit, such as encrypted metadata information or invisible watermarks in content files. The new Chinese regulation requires both types of labels, making it easier to identify AI-generated content.
Explicit Labels
Watermarks on Images: Visible marks that show an image is AI-generated.,Notification Labels: Conspicuous labels at the start of AI-generated videos or virtual reality scenes.,Morse Code Sounds: Audio clips with the Morse code for “AI” (· – · ·) before or after the content.
Implicit Labels
Metadata Information: Encrypted data in content files that include the initialism “AIGC” and details about the companies involved.,Invisible Watermarks: Hidden marks in content that users won’t notice.
The Challenge of Implementing AI Watermarks
While explicit labels are easier to implement, they can be altered or removed. Implicit labels, on the other hand, require companies to work together and adhere to common rules. This could take years to achieve, according to Sam Gregory, the executive director of Witness, a human rights organization in New York.
Social Media Platforms' Role
The new regulation also holds social media platforms responsible for identifying and labeling AI-generated content. Platforms like Douyin, WeChat, and Weibo will need to examine shared files for implicit labels and AI-generation traces. This is a significant challenge, given the vast amount of content uploaded daily.
“If WeChat or Douyin needs to examine every single photo uploaded to the platform and check if they are generated by AI, that will become a huge burden in terms of workload and technical capabilities for the company,” says Jay Si, a Shanghai-based partner at Zhong Lun Law Firm.
“If WeChat or Douyin needs to examine every single photo uploaded to the platform and check if they are generated by AI, that will become a huge burden in terms of workload and technical capabilities for the company,” says Jay Si, a Shanghai-based partner at Zhong Lun Law Firm.
China vs. the EU AI Act
China’s new regulation goes beyond the European Union’s AI Act, which also requires content labeling. The EU Act focuses on explicit disclosure and machine-readable formats. However, China’s regulation adds the responsibility of screening user-uploaded content for AI, something unique to China’s context and unlikely to be replicated in other countries. For more on how AI is handled in the region, see North Asia: Diverse Models of Structured Governance.
“Chinese policymakers and scholars have said that they've drawn on the EU's Acts as inspiration for things in the past,” says Jeffrey Ding, an assistant professor of Political Science at George Washington University.
“Chinese policymakers and scholars have said that they've drawn on the EU's Acts as inspiration for things in the past,” says Jeffrey Ding, an assistant professor of Political Science at George Washington University.
The Impact on Freedom of Expression
The draft regulation is open for public feedback until October 14. While it aims to combat misinformation and privacy invasion, there are concerns about its impact on freedom of expression. The same tools used to identify AI content could also be used to control what users post online.
“The big underlying human rights challenge is to be sure that these approaches don't further compromise privacy or free expression,” says Gregory.
“The big underlying human rights challenge is to be sure that these approaches don't further compromise privacy or free expression,” says Gregory.
The Future of AI Regulation
China’s proactive stance on AI regulation could influence global standards. With the speed and proactiveness of its AI legislation, China is hoping to shape global AI regulation. This approach contrasts with the principles-led governance seen in Japan.
“China is definitely ahead of both the EU and the United States in content moderation of AI, partly driven by the government’s demand to ensure political alignment in chatbot services,” says Angela Zhang, a law professor at the University of Southern California studying Chinese tech regulations. For a deeper dive into China's AI policy, refer to the Stanford University's DigiChina project.
“China is definitely ahead of both the EU and the United States in content moderation of AI, partly driven by the government’s demand to ensure political alignment in chatbot services,” says Angela Zhang, a law professor at the University of Southern California studying Chinese tech regulations.
Comment and Share:
What do you think about China’s new AI regulation? Do you believe it will help combat misinformation or pose a threat to freedom of expression? Share your thoughts and experiences with AI and AGI technologies in the comments below. Don’t forget to Subscribe to our newsletter for updates on AI and AGI developments.
Latest Comments (4)
the idea of mandatory "AIGC" metadata being part of a global standard is but also fraught. from our experience in Pakistan, access to tools that can even read or verify complex implicit watermarks is not universal. how do we ensure this doesn't create another digital divide, where only those with the latest tech can truly discern content origins? it's not just about compliance for big platforms, but about equitable access to information integrity for everyone.
This aligns closely with the UK's AI Safety Institute's focus on provable AI system safety. The idea of mandatory watermarking, both explicit and implicit, for AIGC is certainly a critical component for building public trust and ensuring accountability within regulatory frameworks. It's a pragmatic approach to a growing challenge.
the metadata information part for implicit labels is the real challenge. trying to standardize that across different AI models and platforms later needs serious coordination. gonna cost a lot to implement and maintain that kind of infra.
man this whole explicit/implicit watermark thing we punted on it so hard at my last gig the product team was all about visible labels for "transparency" but eng was like, good luck getting a proper spec. and the implicit stuff, forget about it. you need every platform on board or it's just noise. felt like we spent more time arguing about whether a tiny 'AI' in the corner was "conspicuous" enough than actually building anything. china's just gonna enforce it with a big stick, makes sense for them.
Leave a Comment