Google is now requiring people who create content on YouTube to label any content that was created by AI tools that are deemed as realistic. In a blog post today, Google revealed that the new AI labeling features are now available as a tool in YouTube"s Creator Studio. The company first announced its AI labeling plans back in November 2023.
The new tool is labeled as "Altered content" within the Creator Studio. It simply asks YouTube video makers to say "Yes" or "No" if their video has one or more of the following bullet points:
- Makes a real person appear to say or do something they didn"t say or do
- Alters footage of a real event or place
- Generates a realistic-looking scene that didn"t actually occur
The blog post states that YouTube creators don"t have to label content made by generative AI tools that do not look realistic. One example was "someone riding a unicorn through a fantastical world." The same applies to the use of AI tools that simply make color or lighting changes to videos, along with effects like background blur and beauty video filters.
These new AI-based description labels will appear first on YouTube"s phone apps, and will later expand to the desktop and TV apps. YouTube plans to use "enforcement measures" if a YouTube creator makes a habit of not correctly labeling their realistic AI videos but does not state what they will be like (video removal, demonetization, or suspending an entire account). YouTube also reserves the right to put that label on videos themselves if they have "the potential to confuse or mislead people."
Google is also working on a process for YouTube views to send anonymous alerts to YouTube if a video is using realistic-looking AI-made content.