Reseña del libro "Executive Order 14110 on Safe, Secure, and Trustworthy ai (en Inglés)"
For a few vertiginous months in 2023, artificial intelligence stopped being a specialist debate and became a mass cultural event. ChatGPT reached a hundred million users faster than any consumer product in history, deepfaked audio of public figures circulated freely, and laboratory chiefs went before the Senate to ask, remarkably, to be regulated. On October 30, 2023, President Biden answered that moment with the most ambitious technology directive in American history. This Nimble Books edition presents the complete, authoritative text of Executive Order 14110 on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence - the single document in which the United States first attempted to govern frontier AI from end to end. At more than a hundred manuscript pages and over a hundred discrete directives, EO 14110 is a primary source like no other. It opens with eight guiding principles - safety and security, responsible innovation, support for workers, equity and civil rights, consumer protection, privacy, responsible federal use, and global leadership - and translates each into deadlined tasking across nearly every cabinet department. Its operational heart is Section 4.2, which invokes the Korean War-era Defense Production Act to require developers of the most capable dual-use foundation models to report their training activities, the security of their model weights, and the results of safety red-teaming. The order drew the first quantitative line around model scale: any system trained on more than ten-to-the-twenty-sixth operations fell within the reporting regime. It reached the cloud layer with know-your-customer rules, addressed chemical and biological misuse, tasked Commerce and NIST with watermarking and content-provenance standards, created the Chief AI Officer role across government, and built a staggered calendar of thirty-, ninety-, and three-hundred-sixty-five-day clocks coordinated by a new White House Artificial Intelligence Council. The order's significance only deepened after President Trump rescinded it on January 20, 2025. Much of its substance survived its formal repeal: the compute-threshold concept entered the global regulatory vocabulary and shaped the EU AI Act's systemic-risk trigger, while agency AI inventories and talent pathways took root in the bureaucracy. The order also crystallized an unresolved question - whether sweeping technology governance should rest on a presidential signature the next president can erase, or on the slower durability of legislation. That tension between speed and legitimacy is its most enduring legacy. This Nimble Books print edition reproduces the order faithfully and surrounds it with a full editorial apparatus newly prepared for this volume. A Historical Context essay reconstructs the policy moment of late 2023, the order's place in the arc from EO 13960 through the Bletchley Declaration, and the legal controversy over stretching the Defense Production Act into peacetime. Two complementary abstracts - a plain-language summary that explains the order in everyday terms, and an advanced analytical summary that separates the genuinely enforceable provisions from the merely aspirational - orient readers of every level. A Glossary defines the technical and legal vocabulary; Indexes of persons, places, and concepts make the dense text navigable; and an original computational RKHS knowledge-graph analysis situates the order among related works in technology law and AI policy. For scholars of technology governance, students of administrative law, librarians building reference collections, policy professionals, and any reader seeking to understand how a great power first tried to govern a transformative technology in a single stroke, this is the indispensable edition of a document that will be studied long after its repeal.
