Scammers use large language models (LLMs) to create phishing pages and leave artifacts in texts and tags, like the phrase "As an AI language model...".