Canadian media companies are demanding that OpenAI pay millions in copyright damages, but admit they have no actual knowledge of when, where, or how their data was accessed, scraped or copied.