BBC Verify’s investigation found that Elon Musk’s Grok AI chatbot was incorrectly authenticating AI-generated photos and videos from the Iran conflict as genuine images.