Arthur T Knackerbracket has processed the following story:
The Chinese Communist Party’s (CCP's) national internet censor just announced that all AI-generated content will be required to have labels that are explicitly seen or heard by its audience and embedded in metadata. The Cyberspace Administration of China (CAC) just released the transcript for the media questions and answers (akin to an FAQ) on its Measures for the Identification of Artificial Intelligence Generated and Synthetic Content [machine translated]. We saw the first signs of this policy move last September when the CAC's draft plans emerged.
This regulation takes effect on September 1, 2025, and will compel all service providers (i.e., AI LLMs) to “add explicit labels to generated and synthesized content.” The directive includes all types of data: text, images, videos, audio, and even virtual scenes. Aside from that, it also orders app stores to verify whether the apps they host follow the regulations.
Users will still be able to ask for unlabeled AI-generated content for “social concerns and industrial needs.” However, the generating app must reiterate this requirement to the user and also log the information to make it easier to trace. The responsibility of adding the AI-generated label and metadata falls on the shoulders of this end-user person or entity.
The CAC also outlaws the malicious removal, tampering, forgery, or concealment of these AI labels, including the provision of tools that will help carry out these acts. Although this obviously means that you’re prohibited from deleting the AI label and metadata on AI-generated content, it also prohibits the addition of this identifier for human-created data.
The CCP, through the CAC, aims to control the spread of disinformation and prevent internet users from being confused by AI-generated content via the application of this law. At the moment, we haven’t seen any prescribed punishments for violators, but there is always the threat of legal action from the Chinese government.
This isn’t the first law that attempts to control the development and use of AI technologies, and the EU enacted its Artificial Intelligence Act in 2024. Many may react negatively to this move by the CAC, especially as it’s known for administering the Great Firewall of China to limit and control the internet within China’s borders. Nevertheless, this move will help reduce misinformation from anyone and everyone, especially as AI LLMs become more advanced. By ensuring that artificially generated content is marked clearly, people could more easily determine if they’re looking at or listening to a real event or something conjured by a machine on some server farm.
(Score: 4, Interesting) by JoeMerchant on Saturday March 22, @01:09PM (1 child)
What's the definition of "AI generated"?
If I use spell-check, isn't that AI?
If a writing tool only prompts me one word at a time - is that AI?
For decades now, photo manipulation tools have had "clone" functions which enable very convincing removal of things from images, replacing them with a background that looks normal. Is that AI?
How about a tool like https://hemingwayapp.com/ [hemingwayapp.com] that critiques your writing after you have done it? How much of that AI advice can you accept before you're "letting the tool write for you"?
Cell phone cameras have been "color enhancing" scenes for a long time now, making them look unrealistically colorful and appealing - particularly vacation spots. Many of those algorithms are intelligent artificial enhancement. On-phone portrait enhancers?
What about when Grok "waters itself down" to be a non-AI tool (whatever that means)? How far does it have to go to be non-AI? Can it just be relabeled?
🌻🌻🌻 [google.com]
(Score: 2, Informative) by khallow on Tuesday March 25, @01:46AM
You will know what it is when they punish you for it.
(Score: 4, Insightful) by Booga1 on Saturday March 22, @04:20PM (1 child)
I wonder what the primary driver of this is. Is it because AI feeding on itself degrades things so badly that this is a solution? Is it because the flood of AI trash is going to make it impossible to find anything real?
Motivations aside, this is quite an ambitious law. It's going to have the usual issues with marking the material and making sure it stays marked(similar to DRM). Another issue is the smaller the bit of content gets, the harder it gets to apply the labels.
I wonder what implementation requirements will be.
Is every reply from ChatGPT going to start something like,
How small of an excerpt from a LLM conversation needs a label? If you wanted to publish something you got from ChatGPT, would you have to include metadata to point back to that original session, or is a simple "Original content by ChatGPT" notification enough?
(Score: 1) by khallow on Tuesday March 25, @01:50AM
I think it's more likely that they're worried about automated spreading of information and propaganda. I'm thinking not just a thousand weird ways to use Winnie the Pooh, but also automated ways to route around government censorship.
(Score: 3, Touché) by mcgrew on Sunday March 23, @04:56PM
And Donald Trump will make America great again. With Elon's chain saw.
You really believe the Chinese?
Impeach Donald Saruman and his sidekick Elon Sauron