Instagram’s new AI image tool causes privacy worries

Meta, the company that owns Instagram, has faced criticism from privacy experts because its new artificial intelligence (AI) tool can create images of people with public profiles by default. This tool is called Muse Image AI.

The tool, which was released on Tuesday, allows users to create pictures using faces of people who are featured in public social media posts. Instagram users are not told when their photos are used by this new AI system, which Meta describes as its most advanced image generator yet.

For people who are concerned about their photos being used by Muse Image, the simplest and most effective solution is to change their Instagram account from public to private. This stops strangers from using their public profile pictures as material for the AI.

However, for those who prefer to keep their accounts public, there is another way to manage this. A spokesperson for Meta stated that Muse Image was developed with safety features from the beginning. They explained that private accounts and accounts belonging to users under 18 are automatically excluded. Adult users with public accounts have easy-to-use controls to opt out.

Meta has provided instructions for Instagram users who wish to prevent their photos from being used by the AI image generator. They need to go to the settings section for sharing and reusing content and turn off a specific option. Privacy advocates argue that it is unfair to expect users to find and turn off this feature themselves, describing the process of finding the setting as difficult.

Privacy-focused companies like Proton have pointed out that data sharing is enabled by default. They also mention that the option to disable it is hidden deep within the settings, and often it is public complaints that make users aware of what is happening with their content. They advise users to check these settings carefully because the on and off switches look very similar, making it easy to accidentally leave the feature active.

Meta has confirmed that Muse Image does not allow users to tag accounts of people under 18, and teens cannot use the feature to tag or reference others. However, it remains unclear if children who appear in photos on public adult accounts can be used by the AI. Privacy groups are concerned that children in public photos risk having their likenesses used without permission.

Experts from the Electronic Frontier Foundation believe that this setting should require users to actively choose to enable it (opt-in), rather than it being on by default. They argue that this is a new way photos are being used, and it is not something users likely considered when they first joined Instagram.

Meta’s announcement about Muse Image showed examples of its use, such as creating fashion looks, shopping for clothes, and improving image quality. Another example demonstrated its ability to combine multiple images, like a specific person, a type of bicycle, and an artistic style, to create very detailed and customized results.

Privacy advocates suggest that these creative uses could have negative consequences. While Meta says Muse Image is intended to make AI image generation more personal by allowing users to reference public Instagram accounts, this could be worrying when anyone can use someone else’s account for their creations.

Meta claims Muse Image has strong protections against generating content that breaks its rules. Users can report any AI-generated image they find unacceptable by holding down on the image and selecting a reporting option. The tool is currently available in the Meta AI app, Instagram Stories in the US, and WhatsApp in some countries, with plans to expand to Facebook and add video features.