Internal documents from Microsoft and OpenAI have been revealed, showing the tech giants’ concerns over the use of news content to train AI. Microsoft’s Brent Hecht warned that scraping news was akin to 'the largest theft of labor in human history,' suggesting the practice threatens the very foundations of the news industry.

These revelations come as news organisations accuse the AI firms of violating copyright laws, using large amounts of their content without permission. Nick Turley from OpenAI warned that commercial products like ChatGPT pose an 'existential threat' to publishers, potentially substituting their services.

The documents hint at a 'doom loop' where the AI’s performance is jeopardised by its own content sourcing, a scenario that could have far-reaching consequences for the web and the news industry alike.

The debate over fair use versus exploitation in AI training is now more heated than ever, with these documents providing a glimpse into the tech giants' internal fears and strategies.