It is copyrighted. There's ongoing lawsuits about whether AI training counts as copyright infringement; e.g. Thomson Reuters v. ROSS Intelligence https://www.debevoise.com/insights/publications/2025/02/an-e...

However, they could just do an end run round this. In the UK they're planning to get the government to help them just grab everything for free: https://www.gov.uk/government/consultations/copyright-and-ar...