Meta ordered to remove AI Deepfakes targeting women as Oversight Board slams safeguards

Washington: Meta has moved to remove two AI-generated videos targeting women after the company’s Oversight Board criticised its safeguards against the spread of deepfakes as “fundamentally inadequate”.

The Oversight Board, a quasi-independent body funded by Meta to review its content moderation decisions, on Thursday directed the company to remove the fabricated videos from its platforms.

The board said the cases exposed shortcomings in Meta’s policies on hate speech, misinformation and harassment.

AI videos targeted women

One of the videos falsely portrayed a Labour Party councillor in Scotland making inflammatory comments about refugees.

There was no public record of the politician making the remarks, while the video showed signs of manipulation, including audio that did not match her facial movements.

The second AI-generated video targeted a Muslim woman campaigner in Europe. It falsely portrayed her eating junk food while promoting mindful eating and performing exaggerated exercises, apparently aimed at undermining her work as a health advocate.

The Oversight Board said Meta’s policies remain “consistently and fundamentally inadequate” to deal with the growing volume of AI-generated content on Facebook and Instagram.
Also read | Meta agrees to report child sexual abuse cases in India; Centre warns platforms of action

Board calls for stronger deepfake rules

Oversight Board Co-Chair Pamela San Martin said AI-generated deepfakes were increasingly being used to harass and silence women who participate in public life.

She called for stronger policies, including clearer labels for deceptive AI-generated content, tougher penalties for repeat offenders and the removal of deepfakes targeting private individuals.

The board’s decisions on individual cases are binding on Meta, although its broader policy recommendations are not.

Meta said it welcomed the board’s decision in both cases and would comply with the removal orders within seven days.
Also read | Meta’s new AI ‘Muse’ can manage your accounts, but Sentinel needs your OK for every purchase

Board raises concerns over Community Notes

The latest criticism comes a week after the board urged Meta not to replace professional fact-checking with Community Notes, a crowdsourced system for identifying and debunking false information.

Meta has been testing Community Notes in 16 Spanish-speaking countries in Latin America as part of its move towards replacing its fact-checking programme.

(With AFP inputs)