Meta's Oversight Board investigates explicit AI-generated images on Instagram & Facebook

Following system failures, Meta's Oversight Board has conducted two separate investigations on Instagram in India and Facebook in the US resulting in the removal of AI-generated images of public figures on both platforms.

In recent years, some generative AI tools have expanded to enable users to generate pornographic content, although not all tools offer this capability. Deepfakes have emerged as a concern, particularly in regions like India, where data suggests that women are more frequently targeted in deepfaked videos. While India has contemplated introducing specific regulations addressing deep fakes, no definitive measures have been enacted yet.

Although the country has legal provisions for reporting online gender-based violence, experts observe that the process can be cumbersome, and victims often receive inadequate support. Currently, only a limited number of laws worldwide specifically address the creation and dissemination of pornographic content generated using AI tools. Some U.S. states have enacted legislation targeting deepfakes. Recently, the United Kingdom passed a law criminalizing the production of sexually explicit imagery generated with AI technology.

However, the Oversight Board has invited public feedback, with a deadline of April 30, regarding issues related to harms caused by deepfake porn, contextual insights into the spread of such content in regions like the U.S. and India, and potential shortcomings of Meta's approach in detecting AI-generated explicit imagery. Following an investigation into the cases and public input, the board will publish its decisions on the website in the coming weeks.

These cases underscore the ongoing challenges faced by major platforms in adapting older moderation processes to the rapid proliferation of content facilitated by AI-powered tools. Companies like Meta are exploring AI-driven solutions for content generation, although there are ongoing efforts to enhance detection mechanisms for such imagery.

However, despite these efforts, perpetrators continue to find ways to circumvent detection systems and share problematic content on social media platforms.