Meta takes new approach in labelling AI-generated content

Meta has changed its ‘Made with AI’ label to ‘AI info’, which users can click to see if something is AI generated or modified with retouching tools.

Social media giant Meta has announced changes to how it labels AI-generated content to cater to a wider variety of content – such as photos enhanced by AI.

The company announced changes to how it handles manipulated media earlier this year, as a way to reflect the “broader range of content” that exists on its platforms. This led to a ‘Made with AI’ label appearing on video, audio and images that Meta’s systems detected as AI generated.

But there are reports that this caused confusion for some users, as real content that was brushed up with digital tools was getting labelled as ‘Made with AI’. In an updated blogpost, Meta said its labels were not always “aligned with people’s expectations and didn’t always provide enough context”.

“For example, some content that included minor modifications using AI, such as retouching tools, included industry standard indicators that were then labeled ‘Made with AI’,” Meta said. “While we work with companies across the industry to improve the process so our labelling approach better matches our intent, we’re updating the ‘Made with AI’ label to ‘AI info’ across our apps.”

Meta said users will be able to click this ‘AI info’ label to get more information about how exactly AI was used in the content – to get clarity on whether an image is AI generated or if it was enhanced by AI.

Based on the earlier labelling update, this will continue to be created through Meta’s own AI detection tools and people “self-disclosing that they’re uploading AI-generated content”.

Meta has been focused on AI for some time now and faced severe criticism in the EU in recent months, when it was revealed that the company planned to use customer data to train its AI models.

Noyb, the privacy advocacy group, filed complaints against Meta to 11 data protection authorities over this issue, including Ireland’s Data Protection Commission (DPC). Meta’s plans also caused concerns among many artists using Instagram, who fled to Cara as an alternative site to publish their content.

In response to this pressure, Meta paused its plans to train its large language models using public content on its platforms. The company said this was done at the request of the DPC and called the move a “step backwards for European innovation”.

Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.