Researchers identified more than 50 ads on Meta platforms that violated policies on child sexual abuse material, with AI-generated CSAM appearing in ads.