When filings unsealed in January 2026 copyright lawsuit against Anthropic revealed an internal plan to “destructively scan all the books in the world,” just about everyone on the internet wondered whether the bookless future predicted in dystopian works of science fiction like Fahrenheit 451 would arrive not as piles of books stacked high and torched in some far-off authoritarian regime, but as the mundane expansion of AI training datasets in warehouses across America.
Anthropic called this pursuit “Project Panama,” a program to quietly spend millions on books, slice off their spines with hydraulic cutters, scan their pages, and discard the husks of human knowledge in a quest for new and better training data to help AI learn “how to write well” instead of spewing “low-quality internet speak.” “We don’t want it to be known that we are working on this,” read the memo.
