Is there any evidence OpenAI has been ignoring robots.txt for scraping purposes? AFAIK the main sources of that traffic are still unknown.