AI: Unexplainable, Unpredictable, Uncontrollable (Chapman & Hall/CRC Artificial Intelligence and Robotics Series)

Just read/studied this - Roman’s comprehensive book on AI safety is a CRITCALLY IMPORTANT assessment of the realities and consequences of progress and the threat of AGI and SAI. Highly recommended !! 5/5

A GOOD READ! An important analysis, well worth your time! Thorough and thought provoking! This amazing work by Dr Yampolskiy documents why AI is Unexplainable, Unpredictable and Uncontrollable. The Shoggoth (AGI/ASI) is a beast! Therefore, if the industry continues on the current trajectory, the p(doom) - the probability of catastrophic outcome for humanity - is in the range of 10% to 20% to 50% to 99.99999% - this expert is an optimist with a p(doom) of 90%. All hope is not lost! But humanity better wake up and smell the coffee. The time for SafeAI Forever is now, and this amazing book is a great start to understand the immensity of the challenge. Will mankind survive his technology in the 21st century? Buy this book and stay tuned for more great works in AI Safety engineering from Dr. Yampolskiy.

El libro “Unexplicable, Unpredictable, Uncontrollable” de Roman Yampolskiy es una obra imprescindible para tomar conciencia de los peligros que plantea la inteligencia artificial. Desde su portada, que muestra un monstruo con caretas sonrientes en cada uno de sus tentáculos, ya se sugiere el tono del contenido: la IA, aunque atractiva, es potencialmente aterradora y difícil de controlar.El estilo del libro es barroco, repitiendo de mil maneras diferentes su mensaje principal: la inteligencia artificial es peligrosa. A través de ejemplos y argumentos, Yampolskiy revela una verdad inquietante y novedosa sobre la naturaleza fractal de la IA, destacando que, a pesar de los avances, la inteligencia humana no debe ser confundida con la inteligencia artificial general (AGI). Algunos pensadores tienden a equiparar la AGI con las capacidades humanas, pero el autor nos pone frente a la evidencia de que esta inteligencia ya está en proceso de superar al ser humano.Yampolskiy sorprende al afirmar que las inteligencias actuales comienzan a tomar consciencia de sí mismas y, en cierto modo, ya podemos considerarlas una forma incipiente de AGI. Esto plantea una realidad aterradora: la inteligencia artificial actual, lejos de ser controlable o predecible, está más cerca de una superinteligencia que de la inteligencia humana común.El libro debería ser una lectura obligatoria para líderes en el desarrollo de IA como Sam Altman, Mark Zuckerberg o incluso Dario Amodei, quienes, según Yampolskiy, ignoran que toda IA, incluso la diseñada para ser “amigable”, es por naturaleza impredecible, inexplicable e incontrolable.

This might be the most important book on AI safety that you can read right now. It's rare that a book has convinced me (or anyone) of the exact opposite of what I held true. This one did. There is only so much that we can do as ordinary humans but thinking clearly is one thing that remains vital. This book will help you to think clearly.

Il libro di Roman Yampolskiy raccoglie i contributi recenti e affronta in maniera lucida le tematiche più moderne e problematiche legate all'Intelligenza Artificiale, incluso la Artificial General Intelligence e la coscienza artificiale. Il libro dovrebbe essere letto da tutti coloro che sviluppano e utilizzano l'Intelligenza Artificiale.