An amended lawsuit alleges Grok generated about 7,000 AI-created child abuse images and says xAI failed to stop the activity or provide key evidence.