Meta announced in February that it would begin labeling some AI-creations photos on its social platforms. Since May, Meta frequently has been annotating some photos with a "Made with AI" tag across its Facebook, Instagram, and Threads applications.

Still, it has enraged users and photographers as the company labels photos with the "Made with AI" tag even when they haven't been created using AI tools at all.

It brings with it hundreds of examples of Meta automatically applying the tag to photos that were not created via AI. There is, for example, this photo of Kolkata Knight Riders clinching the Indian Premier League Cricket tournament. Significantly, the tag is available only on the mobile apps and not on the web.

Many other photographers have complained that their images have been accidentally branded with the "Made with AI" label. They stress the fact that a tag should not be associated with the process of only editing a photo within a software.

Former White House photographer Pete Souza posted on Instagram one of his photos labeled. According to Souza, in an email to TechCrunch, Adobe changed the cropping tool functionality, and apparently "flatten the image" has to be done before saving it as JPEG image. But Souza is suspecting this particular action triggers Meta's algorithm in attaching this label.

"What's frustrating is that the post made me include the 'Made with AI' even though I unchecked it," Souza said to TechCrunch.

Meta won't publicly comment on TechCrunch's questions regarding Souza's experience or other photographers' posts that said their posts were tagged wrongly. However, within hours of the publication of this story, Meta responded that the company is reviewing its practice of using labels to indicate the amount of AI that went into an image.

"Our intention has always been to enable people to know when they see content that's been created with AI. We're considering recent feedback, continuing to assess our approach so our labels capture the degree of AI in an image," a Meta spokesperson said in an email to TechCrunch.

Meta said earlier this year in a February blog post that it detects the label using the metadata of images.

It said that at that time, "we're building industry-leading tools to identify invisible markers at scale—namely, the "AI-generated" information within the C2PA and IPTC technical standards—so we can apply labels to images from Google, OpenAI, Microsoft, Adobe, Midjourney, and Shutterstock as they roll out their plans for image metadata.".

PetaPixel made public last week that Meta has started attaching the 'Made with AI' tag whenever photographers use tools such as Adobe's Generative AI Fill to remove objects.

While Meta hasn't explained when it typically applies the label, some photographers have taken Meta's side arguing that any use of AI tools should be disclosed. It said it is actually working actively with the companies that have AI-powered tools for creation in order to refine its approach.

"We're relying on industry-standard indicators that other companies have in content from their tools, so we're actually working with these companies actively to make the process better so our labeling approach matches our intent,"

For now, Meta attaches no special flags to indicate whether a photographer used a tool clean up their photo, or created it with AI. It may be tricky for users to discern just how much AI is at play in any given photo. Meta's label does indicate that "Generative AI may have been used to create or edit content in this post" — if you tap on the label.

Despite this approach, there are still plenty of pictures on Meta's platforms which are widely recognized as purely AI-generated products and Meta's algorithm hasn't labelled it. In the next few months, U.S. elections will be held, and so, social media companies are under more pressure than ever to correctly handle AI-generated content.