Facebook’s AI wants to scan your photos — even those you’ve never shared

Facebook is asking users for permission to scan their phone’s camera roll, including photos the app was previously never granted access to, so it can generate versions edited with artificial intelligence. The new feature is raising serious questions, given the platform’s opaque terms of service regarding its collection of AI-generated data.

Users attempting to create a new story on Facebook are being met with a pop-up that asks whether they’ll permit “cloud processing,” an option that allows the app to upload photos from their phone “on an ongoing basis.” The function has not yet been rolled out to all Facebook users.

“The best of your camera roll, curated for you: Get ideas like collages, recaps, AI restyling or themes like birthdays or graduations,” the pop-up states.

Data retention

Facebook said it will choose and upload photos based on information such as time, location, themes, as well as the people and objects present.

The app said the photos it accesses won’t be used for ad targeting but are subject to the AI terms and services of Meta, Facebook’s parent company. Straight Arrow News reached out to Meta on Friday, June 27, to inquire about the new feature but did not receive a reply.

As reported by TechCrunch, the terms state that any image processed by Meta’s AI can be analyzed “to offer innovative new features, including the ability to summarize image contents, modify images and generate new content based on the image.”

The terms also state that Meta can retain and use any personal information shared during the creation of AI content, such as prompts, feedback and “other content.”

Aside from the pop-up, the feature is also listed under Facebook’s settings in the “preferences” section. By visiting the “camera roll sharing suggestions” page, users can choose whether to toggle on or keep off the cloud processing function.

AI controversy

Although the feature has received little pushback so far, it wouldn’t be the first controversial move from the tech company regarding AI. Meta was recently sued by a group of authors who accused the tech company of breaching copyright by training its AI on their books.

A judge ultimately ruled in Meta’s favor, stating, according to The Guardian, that the authors failed to present sufficient evidence that Meta’s AI would cause “market dilution” by releasing content similar to theirs.