Meta has released a new artificial-intelligence image tool that can generate pictures based on the photos of people with public Instagram accounts, and the fact that it is turned on by default has drawn criticism from privacy campaigners and some users.
What the tool does
The feature, part of a new image generator Meta has built into its apps, lets someone reference a public Instagram account in a prompt so that the resulting AI image draws on that person's photos, the BBC reported. It is enabled by default: people with public profiles are included unless they go into their settings and turn the option off. According to the reporting, users are not necessarily notified when their images have been used this way.
Why critics are alarmed
Digital rights groups say the design gets consent backwards. The campaign group Foxglove described the approach as a recipe for trouble, and other privacy advocates argued that it treats people's photographs as material to be used without asking first, TechCrunch reported. The core objections are familiar ones in debates over AI and personal data: whether people have agreed, whether they are told, and whether they can meaningfully control how their image is used.
Meta's position
Meta says users are in control because the setting can be switched off, and it points to the tool's intended creative uses. Critics counter that an opt-out, rather than an opt-in, puts the burden on individuals who may not know the feature exists. Newsparlor has not independently tested the tool or verified individual cases.
The wider context
The launch comes as Meta's handling of personal data continues to draw scrutiny. The company remains bound by a 2019 settlement with the US Federal Trade Commission, which included a $5 billion penalty, over earlier privacy failures, and its use of user data to build AI systems has prompted questions from regulators in Europe about compliance with data-protection law. Whether authorities examine this particular feature is not yet clear. For now, it sharpens a running tension between the appetite of AI companies for large amounts of real-world data and the expectations of the people that data describes.
