Amid criticism, Meta reins in new AI tool that automatically accessed public Instagram images

So Meta has 'reined in' its Muse Image AI tool after it was caught automatically helping itself to public Instagram photos as training data. How generous of them to retreat from a feature nobody asked for, after the backlash made it impossible to ignore. The company's statement will no doubt frame this as a thoughtful course correction, but let's be honest: they only pulled back because the PR cost exceeded the data haul.

This is the same Meta that has spent years insisting that anything posted to a public Instagram account is fair game for their AI ambitions. The Muse tool didn't just scrape images — it used them as 'references,' which is a lovely euphemism for 'we took your holiday snaps and fed them into our machine without a by-your-leave.' The fact that they thought this would slide past a public still smarting from Cambridge Analytica shows either breathtaking hubris or a complete disconnect from reality.

The regulatory gap here is a chasm. There's no US federal privacy law that clearly prohibits this behaviour, so Meta's self-correction is voluntary and reversible. They'll wait until the heat dies down, then quietly re-enable the feature with a slightly different toggle. And the cycle continues. The only surprise is that they bothered to pretend to care this time.