Meta investigates AI-generated Instagram accounts accused of sexualising disabled people

Meta, the parent company of Instagram, has launched an investigation into AI-generated accounts on its platform that appear to sexualise and exploit images of disabled people. The issue was highlighted after the BBC identified dozens of profiles sharing artificially generated images and videos of women depicted with disabilities, including Down’s syndrome, vitiligo, missing limbs and visible scarring. Many of these images portray women in revealing clothing or sexualised poses. Some of the accounts have rapidly attracted large audiences, with one profile claiming to represent conjoined twins amassing around 400,000 followers within months of being created.

Disability rights organisations and medical charities have strongly condemned the trend. Kamran Mallick, chief executive of Disability Rights UK, described the accounts as horrific and accused them of stripping disabled people of dignity and agency for profit and titillation. Gemini Untwined, a charity supporting children born conjoined at the head, called the portrayal of conjoined twins as entertainment morally reprehensible.

Experts say the proliferation of low-cost generative AI tools has made it easier to create such content. Dr Amy Gaeta of the University of Cambridge noted that biased training data can result in hypersexualised depictions of disabled people, even without explicit prompts. While some AI tools impose restrictions on explicit content, others can be bypassed.

Regulators including Ofcom are monitoring the evolving risks linked to AI, and the Online Safety Act requires platforms to enforce their terms consistently and protect users from harmful or abusive material. Meta has stated it removes content promoting sexual exploitation or attacks based on protected characteristics and is reviewing the flagged accounts.