Meta's Facebook, Instagram Ran AI 'Nudify' Ads in Policy Breach

According to Bloomberg, a report by the nonprofit Tech Transparency Project found that Meta Platforms Inc.'s (META) Facebook and Instagram served thousands of ads for AI "nudify" apps that strip clothing from images, violating the company's policies. The ads were delivered by one of Meta's few advertising partners in China.

Details of the Violation

The report highlights that despite Meta's explicit ban on sexual content, pornography, and apps that undress images, the ads continued to appear. This raises questions about the effectiveness of Meta's content review systems, which rely on a combination of AI and human moderators.

Company Response

Meta has not yet issued an official comment on the report. However, the company has previously stated its commitment to strict policies against unwanted sexual content and uses automated and manual reviews to enforce them.

Precedents and Context

This is not the first time Meta has faced criticism for policy-violating ads. The company has been scrutinized for allowing misleading or harmful ads, especially during election periods. Reliance on external ad partners, particularly in China, adds complexity to content moderation.

Potential Financial Impact

This disclosure could lead to increased regulatory scrutiny from authorities in the EU and US, potentially resulting in fines or corrective actions. It may also erode advertiser and user trust in Meta's ability to enforce its policies.