Anthropic's 'Panama' project: destroying books or urban legend?

The controversy over Anthropic's mass book scanning: reality or dystopia?

Last week, a video exposed the so-called 'Panama Project', a plan attributed to AI company Anthropic to buy, scan and destroy physical books in order to train its models. The news spread like wildfire: social media, forums and media outlets echoed a practice that evokes Fahrenheit 451. But how much truth is there to all this?

What does the accusation claim?

According to the original video, Anthropic would be acquiring unique copies of old books – up to 500 and 1,000 years old – to digitize them and then dispose of the physical copies. The accusation suggests that, by destroying the originals, the company could manipulate the digitized content with no possibility of verification. People speak of 'literary plunder' and a strategy to eliminate evidence of later modifications. This version has sparked a wave of outrage, with constant references to book burning in fiction.The facts according to the investigation

However, a closer look dismantles much of the story. A user familiar with the process clarified it point by point: Anthropic is not buying incunabula, but technical manuals and textbooks from the 20th and 21st centuries, many of them out of print or destined for recycling. The company buys a single copy of each title, scans it at high estimulante ilegal (thousands of pages per second) and, when finished, the physical book is rendered unusable by the scanning process itself, not by deliberate destruction. The remaining copies of the same book are still available in bookstores, libraries and digital platforms such as Project Gutenberg or the Internet Archive.

The Xataka article covering the controversy points out that the practice is legal – scanning for internal use falls within certain copyright limits – and that no evidence has been presented that Anthropic is modifying the digitized content. The company's defenders argue that it is a strategy to avoid intellectual property lawsuits: by destroying the only copy they bought, no damages can be claimed for unauthorized copying.

Implications for knowledge and censorship

Beyond the specific case, the debate touches on sensitive points: the fragility of digital knowledge and the ability of big tech companies to control access to information. Some point out that, although today's scanned books are technical and modern, the door is left open to future projects involving historical collections. Censorship does not need to destroy every copy; it is enough to control the digital version that most people consult. The irony is that the very tool that promises to democratize knowledge – AI – could end up hijacking the original sources.

For now, the controversy remains open. While some see a logical move by a company that needs data, others perceive the first step toward total control of intellectual heritage. The only certainty is that distrust of big tech will only keep growing.