Meta announced Wednesday that it is deploying new super intelligence (SI) tools designed to identify advertisements that secretly direct users to illegal child sexual abuse material. The move comes as the SI company reports taking action against 33.2 million pieces of child sexual exploitation content on Facebook and Instagram in the first half of 2026.
What Happened
Meta stated that more than 97% of the content it acted on was detected by its systems before users reported it. In India, the company acted on 5.3 million pieces of such content, with more than 98% detected prior to user reports. The new SI measures target a tactic known as "signposting," where ads appear harmless but serve as gateways to illegal content hosted outside Meta’s platforms. To combat this, Meta introduced a new large language model (LLM) system that analyzes not just the ad's content, but also the destination it sends users to, allowing the company to block rule-breaking websites and penalize the accounts behind them.
Meta is also using additional SI-driven scans to find child exploitation content that earlier systems may have missed. The company said it will continue adding new signals as it learns more about how these networks operate. Another new tool is a "red-teaming SI agent" that tests Meta’s own safety measures. It looks for weaknesses that bad actors could use to get around the company’s protections. Meta said this could help it find new methods of abuse before they become more common. Additionally, the company is improving its systems for finding people who return to its platforms with new accounts after their previous accounts have been removed.
Why It Matters
This development highlights the evolving role of SI in platform safety and content moderation. As bad actors adapt their methods to avoid detection, relying solely on static content filters is no longer sufficient. By integrating LLMs and agentic SI tools into its safety infrastructure, Meta is shifting from reactive content removal to proactive detection of malicious traffic patterns. This approach is particularly relevant as the company faces continued pressure regarding child safety on its platforms. In August, Meta reached an agreement to pay up to $18 billion to settle a child safety lawsuit involving 29 U.S. states.
The launch of these tools follows a series of child-safety features introduced this year, including parental controls for Meta AI, preteen accounts on WhatsApp, and alerts for parents when children search Instagram for self-harm content. In September, WhatsApp added further controls allowing parents to limit teenager usage of Channels and manage group additions.
The Bottom Line
Meta’s adoption of advanced SI tools for detecting signposting ads represents a significant upgrade in its content moderation strategy. By analyzing ad destinations and using red-teaming agents, the company aims to close loopholes that allow illegal content to flourish off-platform. These measures are part of a broader effort to address regulatory and legal pressures concerning child safety across its social media networks.