Meta has taken down 33.2 million pieces of child sexual exploitation content across Facebook and Instagram during the first half of 2026, the company announced Wednesday. Alongside this enforcement effort, Meta is introducing new AI-powered tools designed to identify ads and accounts that may be quietly funneling people toward illegal child abuse material.

According to the company, over 97% of the content it acted on was flagged by its own systems before any user reported it. In India specifically, Meta removed 5.3 million pieces of such content during the same period, with more than 98% caught by automated detection ahead of user reports.

The company explained that its newest AI tools target a specific tactic: ads that appear harmless but are used to redirect users to illegal content hosted beyond Meta's own platforms. To address this, Meta has built a new large language model (LLM) system capable of detecting what it calls "signposting" — advertisements that look ordinary but are suspected of pointing users toward illegal content or other harmful activity elsewhere online.

Meta says this approach represents a tactic it has recently observed bad actors adopting as they keep evolving their methods to dodge detection. Even when the ads themselves don't contain illegal material, they can serve as entry points to websites hosting child sexual abuse material. As a result, Meta says it now examines where an ad directs users, rather than only evaluating the ad's content. This information allows the company to block websites or other destinations that violate its rules and to take action against the accounts operating behind them.

Beyond that, Meta is running additional AI-driven scans to surface child exploitation content that earlier systems might have overlooked. The company stated it will keep incorporating new signals as it gains a deeper understanding of how these networks function.

Another tool in Meta's arsenal is a "red-teaming AI agent" that stress-tests the company's own safety measures. It hunts for vulnerabilities that bad actors could exploit to bypass Meta's protections. According to Meta, this could help the company uncover new abuse methods before they gain wider traction.

In addition, Meta is upgrading its systems for identifying individuals who attempt to return to its platforms by creating new accounts after their previous ones were removed.

These latest measures arrive as Meta continues to face scrutiny over child safety on its platforms. The company has been the target of lawsuits and criticism from lawmakers concerned about the risks its social media services may present to young users. In August, Meta agreed to pay up to $18 billion to settle a child safety lawsuit brought by 29 U.S. states, addressing concerns about protecting children online.

Throughout this year, Meta has rolled out a number of child-safety features, including parental controls for Meta AI, pre-teen accounts on WhatsApp, and alerts that notify parents when children search Instagram for self-harm content. In September, WhatsApp introduced additional parental controls. These allow parents to limit how teenagers use Channels, control who can view their status updates, and determine who can add their children to groups. Parents can also opt in to receive notifications about certain group activity.