Meta has announced new measures aimed at identifying advertisements that may lead users to illegal content, particularly child exploitation material, following criticism over its handling of such issues. A report by the non-governmental organization Tech Transparency Project (TTP) in mid-September revealed significant shortcomings in how Meta's platforms, such as Facebook and Instagram, moderated ads that featured AI-generated images of child exploitation. In response, Meta stated in a report published on October 6 that it had removed 33.2 million instances of child sexual exploitation content across its platforms between January and June 2026. The company is introducing new tools to detect ads that appear harmless but are designed to direct users to illegal content hosted outside of Meta's platforms. "An advertisement may seem innocent on the surface but could be used to subtly guide users to content that violates our policies," Meta explained. A new AI model will be used to identify such deceptive ads, while the company will also conduct deeper analysis of the links included in advertisements. Links to websites associated with illegal activities will be blocked to prevent access. Meta is also enhancing its ability to detect repeat offenders, focusing on individuals who have been banned from the platform but attempt to create new accounts. Additionally, the company is employing AI-driven "red teaming" to test its security measures and identify emerging criminal tactics. This approach involves simulating attacks to uncover vulnerabilities and improve defenses. Meta emphasized that it has long had a team of experts, including former FBI investigators, dedicated to identifying and removing networks of predatory accounts. The company stated that it continues to investigate these networks at a systemic level, aiming to take action against entire groups rather than focusing only on individual accounts.