How to identify AI content on Meta products
Generative artificial intelligence (generative AI) is a tool that helps create or modify content such as images and text in new ways.
How generative AI works
Generative AI models are trained using billions of pieces of information. To generate new content, generative AI models learn about patterns and relationships between these pieces of information based on prompts or instructions.
For example, you might type a request in a generative AI tool asking for an image of 3 brown dogs running through a meadow of flowers. Or you might request a joke written in the style of a historical figure, like Shakespeare.
How AI-generated content is identified on Meta products
Content created or modified using AI tools may be identified and labeled to help promote transparency across Meta products.
There are many AI tools that can create or modify content. Some content that is made or modified using one of these tools may have an industry-standard signal that identifies the content as being created or modified using AI tools. For example, a person might edit an image using a common photo editing tool to share with their followers.
If a photo editing tool uses AI to change the size or color of an image, then the image may have a signal that shows it was edited using AI. These signals are read by Meta’s systems to determine if a label is applied to the content.
When AI-generated content is labeled on Meta products
There are a few ways organic content, or content that is not an ad, will be labeled on Meta products. These can include:
- When Meta’s systems detect AI signals in organic content
- When people label their AI-generated content
Content made with third-party tools may contain signals that show whether the content was created or modified with AI. Content with signals we can identify indicating that it was created with AI will be labeled AI info. Content with signals indicating that it was modified using AI will not have the AI info label directly on the post, story, reel or thread. You can learn more about how the post, story, reel or thread was modified by AI by clicking the 
People using Meta products may also be able, and in some cases required, to label their content with AI info when they share AI-generated content or content that was modified using AI.
Meta may continue to update the approach around transparency as more is learned about how people use AI-generated content across Meta technologies.
Other AI labels you might see
Any content that is created or edited using Meta’s AI tools and shared to Facebook, Instagram, or Threads as a post, story, reel, or thread may be automatically labeled as AI content or in some cases, feature a visible watermark.
Note: Not all AI content contains the information needed to identify it.
When the AI info label may be required
On Facebook, Instagram and Threads, Meta may require an AI label when content has photorealistic video or realistic-sounding audio that was digitally created, modified or altered, including with AI. This requirement does not apply to images. However, images may still receive a label if the systems detect that they were AI-generated or if they were modified using AI.
Note: There may be penalties for content shared without a label when it is required.
Examples of digitally created content that requires a label on Facebook, Instagram and Threads:
- A video that appears realistic of a group of people walking around an outdoor market
- An audio file of two people talking
- A song created using AI-generated vocals
- A reel narrated with a realistic AI-generated voiceover
Examples of digitally created content that does not require a label:
- Video of an outdoor landscape, created in a style resembling a cartoon
- Image of a person riding a bull (Meta does not require a label for images)
Learn more about this requirement in the Transparency Center.
Note: We will update our transparency approach as we learn more about how people use AI-generated content on our platforms. Learn more about generative AI in the Privacy Center.