If you’ve noticed an “AI Info” label appearing on some of our Meta platforms images recently, you might understandably assume that the image was generated by artificial intelligence.
In our case, that’s often not what happened.
Like many digital publishers, designers, and editors, we use tools such as Canva’s background remover and other AI-powered features like Magic Eraser, Magic Grab, Upscaler, and more editing tools when creating our visuals.
They help us remove unwanted objects, clean up backgrounds, reposition existing elements, and make photographs work better within a specific design.
The important part is that we’re usually starting with a real photograph.
We’re taking an existing image and using an AI-powered editing feature to make a specific adjustment to it.
And apparently, Meta doesn’t always make that distinction very obvious to the person looking at the post.
They’re labelling content that is slightly AI-edited with the same label as content that is fully AI-generated.
There is a difference between protecting people from fabricated reality and treating every image that has touched an AI-powered editing tool as though it was fabricated.
Meta has already been expanding its systems for identifying AI-generated and AI-edited content under the “AI info” label.
The company says it uses signals such as industry-standard metadata and other technical indicators to identify content that has been generated or edited with AI tools.
That means something as simple as removing an object from the background can potentially become part of the image’s AI history.
This is where the conversation gets complicated.
There is a genuine difference between “this image was generated by AI” and “AI was used as one of the editing tools used to create this image.”
If we use a photo, for example, of a person at an event and later use an AI-powered tool to remove an unwanted object behind them, the photograph itself is still a real photograph.
The location, person, and original moment exist. AI was used during the editing process.
The same applies to a designer using Canva to remove a background, isolate a subject, or make small visual adjustments. This is literally designing.
At the same time, the technology behind these tools is still generative AI.
We’re not saying it’s wrong for a platform to say that AI was involved somewhere in the creation or editing process. But what’s actually AI-edited and what’s AI-generated needs to be differentiated.
A simple “AI” label can leave out a lot of context.
And this is becoming a much bigger conversation across the creative industry.
The European Union’s AI Act has now brought new transparency requirements for AI-generated and manipulated content into effect, while platforms including Meta are developing their own systems for identifying and labelling AI content.
The goal of these rules is understandable. With realistic AI images, videos, and audio becoming increasingly difficult to distinguish from reality, people deserve to know when they are looking at synthetic content.
That distinction is going to become increasingly important for photographers, designers, publishers, brands, and pretty much anyone creating content online.
For us, transparency is the important part. If an image is AI-generated, we don’t want to present it as a real photograph. But if a real photograph has simply been edited using an AI-powered tool, an “AI” label can tell only part of the story.
This is happening with hundreds of users online now. So, if you’re wondering why your content has that label, you know why.
