Meta is going through a backlash over its new AI software Muse Picture, which might generate photos utilizing different individuals’s profile photos with out telling them.
It’s one in all many text-to-image instruments publicly out there, which because the title suggests can create photos from a number of strains of straightforward written textual content.
Muse Picture is obtainable by way of the Meta AI app and internet browser, in addition to on WhatsApp and in Instagram Tales for US customers.
Whereas Meta says customers can decide out of their picture getting used even with a public account, Donald Campbell, advocacy director at tech justice non-profit Foxglove advised the BBC it was an “apparent recipe for catastrophe”.
“We have already seen a listing of harms from non-consensual AI-altered photos on social platforms simply up to now 12 months,” he stated.
“It’s onerous to see why Mark Zuckerberg thinks facilitating but extra of this creepy picture manipulation is a good suggestion.”
The function is more likely to face heightened scrutiny as regulators and campaigners elevate considerations about AI-generated photos, with Ofcom currently investigating X over Grok’s function in creating and sharing non-consensual AI-altered photos of actual individuals.
Privateness Worldwide additionally criticised the function, telling the BBC it was “the most recent signal AI firms see individuals’s photos and information as uncooked materials to be exploited”.
“Pulling actual customers into generated images with out specific consent is a privateness landmine ready to detonate,” one user wrote on X, external.
Meta stated a devoted setting, separate from account privateness controls, permits customers to decide out even when they’ve a public account.
To take action, customers should go to Instagram’s settings menu, choose “Sharing and Reuse” and swap off “Permit individuals to reuse your content material on Instagram and with AI options at Meta” for posts and reels.
