Meta has launched Muse Image, an AI tool that can generate new pictures from photos on public Instagram and Facebook accounts, and it has drawn immediate criticism.
The controversy centres on one capability rather than the tool as a whole.
By tagging a public account in a prompt, any user can pull that person's photos into an AI-generated image, without the subject's knowledge or approval.
The company owns Instagram, Facebook and WhatsApp, and the feature runs across all three alongside its Meta AI app.
The first objection is consent.
People who posted photos publicly did so to be seen, not to become raw material for image generation by strangers.
Meta neither asks the subject for permission nor notifies them when their likeness is used, so a person may never learn their face has been reworked.
The second objection is the default setting.
Public accounts are opted in automatically, and only private accounts and those belonging to under-18s are excluded.
Users who object must find and switch off a buried setting, which critics say reverses the burden, placing it on the subject rather than the person doing the generating.
One widely shared complaint asked why the feature was not opt-in rather than on by default.
The third and most serious objection concerns misuse.
Making it easy to manipulate real people's images opens the door to harassment, impersonation and non-consensual altered pictures, including sexualised deepfakes.
Foxglove, a tech-justice group, told the BBC the design was an obvious recipe for disaster, citing a year of harm from non-consensual AI images.
Privacy International said it was a further sign that AI companies treat people's images as material to be exploited.
Meta's defence rests on user control and labelling.
The company says people can turn the feature off at any time, and that every output carries an invisible watermark marking it as AI-generated.
Critics counter that an opt-out most users will never see is weak protection, and that watermarks do nothing to stop the image being created or circulated in the first place.
The timing sharpens the risk for Meta.
The tool arrives just as regulators are moving against exactly this kind of capability.
Britain's Ofcom is investigating X over its Grok image generator, and India has built a deepfake takedown regime after pressing X over non-consensual images.
A Meta feature that turns public photos into AI images walks straight into that scrutiny.
The deeper issue is a shifting definition of what public posting implies.
For years, a public photo meant others could view or share it.
Muse Image extends that to mean others can transform it, using a person's likeness in images they never agreed to.
That is the line critics argue Meta has crossed, and why a tool pitched as creative has instead become a privacy flashpoint.
The label, they note, is pointed: the users are the ones being used.
