„Project Panama”: Cum a distrus Anthropic milioane de cărți

Recent court documents reveal that Anthropic secretly initiated a large-scale project to buy, destroy, and scan millions of books for training its AI model, Claude. The internal program, named 'Project Panama', aimed to create a comprehensive database of high-quality texts. To expedite the digitization process, Anthropic outsourced the task to specialized companies that used industrial machines to remove book spines and scan pages. Once digitized, physical copies were discarded and recycled. The project was kept highly confidential, with explicit instructions to avoid public knowledge. In 2024, a group of authors filed a lawsuit against Anthropic, claiming illegal use of copyrighted works for AI training. Anthropic settled for $1.5 billion, marking a significant case in AI copyright disputes. Other companies like Meta and OpenAI also face similar lawsuits from creators.