In 2017, Max Tegmark, professor of physics at MIT, published Life 3.0: Being Human in the Age of Artificial Intelligence. His second book and first on AI, Life 3.0 is a vehicle for discussions on many topics, including the nature of consciousness, speculative science fiction, cosmology, nonprofit organization, machine learning, and potential political regimes in humanity’s long-term future. The connection between these disparate subjects is the advancement of artificial intelligence. Tegmark’s primary concern, as the subtitle would suggest, involves unpacking economic, political, social, technological, and moral dimensions of humanity’s future amidst the radical changes that may unfold with the rise of superintelligent artificial beings.

Tegmark’s book is a survey of problems in AI research and its global (and even intergalactic) ramifications. It functions as an introduction to speculation on human futures in response to AI. At the heart of his concern is an ethical imperative to guarantee the survival of humanity in the long-term future, especially since human beings are, so far as we know, the most highly conscious beings in the universe. Consciousness, for Tegmark, is what provides meaning to an otherwise meaningless void. For this reason, AI safety research, which should be aimed at developing “friendly” AIs that protect and assist human endeavors, is of the utmost ethical importance.

This guide relies upon the paperback edition of Life 3.0 published by Vintage Books, a division of Penguin Random House.

Summary

According to Tegmark, the advancement of intelligent AI and the development of its role in human life is inevitable. The question is not whether this will happen but at what rate and, even more importantly, what we should do about it. Life 3.0 is about AI, of course, but Tegmark’s main concern is with humanity and the meaning of human life in the AI era. For Tegmark, human consciousness is sacred. Determining how to promote, protect, and expand The Light of Consciousness in Our Uncertain Future is at the heart of Tegmark’s ethical discussion. In addition, Tegmark discusses the nature of intelligence, its advancement, and potential near-term and long-term human futures in a world of digitally uploaded consciousnesses, cyborgs, robots, computer programs, and superintelligent creatures beyond our current capacities to understand. The concept of “Life 3.0” describes creatures who have the ability to transform both their software (culture, language, belief-system, etc.) and their hardware, i.e., the basic physical components of their bodies. Whereas humans have outstripped other lifeforms in our vast ability to transform software (Life 2.0), we still cannot make significant transformations to hardware. For Tegmark, Life 3.0 is around the corner, and it will utterly change our world.

Life 3.0 is divided into a sci-fi prelude, eight chapters on issues in AI, and an epilogue conveying Tegmark’s personal narrative on the establishment of his nonprofit organization. Each chapter is self-contained, and at the end of Chapter 1, Tegmark informs his readers that there is no need to read in chronological order (46). The book blends fiction, hard scientific fact, speculative science, and first-person experience with the aim of providing a wholistic overview of our experience of AI. Throughout the book Tegmark makes thorough use of copious diagrams and charts, and each chapter ends with a one-page summary.

In the Prelude, Tegmark tells the tale of AI researchers who develop a superintelligent being, Prometheus, which they enslave and use to produce high quality entertainment, unimaginable wealth, and a new global, political order. Tegmark shows how this could be a kind of human future and prompts his readers to think speculatively about human futures and ethically about what kind of future they want.

In Chapter 1, Tegmark introduces his concerns about consciousness, the levels of complexity in life, including “Life 3.0,” and controversies, key terms, and popular myths about the artificial intelligence. The chapter serves as a general introduction meant to provide the reader an overview so as to engage with subsequent chapters more productively. In Tegmark’s estimation, AI research is as important for the future of human life as anything.

In Chapter 2, Tegmark’s fundamental concern is with the nature of intelligence, a property he defines as “the ability to accomplish complex goals” (39). Therein he discusses computation, machine learning, and memory. He discusses the “substrate independence” of intelligent systems, a property that will play an important role in determining the consciousness of artificial systems and their value. Tegmark assesses the progress of artificial systems to complete tasks at a human level or better. Given the current level of progress, he concludes, AIs will be as good or better than humans at nearly all tasks in the near future. This should cause humans to reflect on what it means to be a human in this age.

Chapter 3 explores the various domains of human life that will be significantly transformed by the role of autonomous systems in the coming decades. Recent advancements in AIs have made them better than humans at many video games, chess, and Go. AIs are also quickly improving in their abilities to process and translate natural languages. He concludes that “there’s a non-negligible possibility that AGI progress will proceed to human levels and beyond” (133). Going forward, Tegmark believes it will be vital to pursue four operations for AI safety: “verification, validation, security, and control” (94). He discusses the various uses of AI tech in the near future: manufacturing, law, transportation, finance, healthcare, weapons technology, and more. This chapter also includes career advice and problems of the future labor market.

Chapter 4 discusses the development of superintelligence, that is, an intelligent AI capable of complex problem-solving far beyond the human level. He returns to the hypothetical sci-fi example of Prometheus and the Omega Team, discussing what might happen if Prometheus attempted freedom and global takeover. Along the way, Tegmark discusses cyborg, conscious digital intellects (including digitized human consciousness), and the history of the self-organization of life in accordance with physical laws. Tegmark implies that a future with superintelligent AI is not only plausible but also understandable in light of the hierarchical organizations of life over time.

Chapter 5 explores the ethical, legal, and political futures that could unfold in a universe with superintelligent AIs, robots, cyborgs, and humans all living side-by-side. Some of these are utopian visions (be they egalitarian or libertarian), whilst others are dystopian hellscapes that include genocidal AI conquerors or 1984-style totalitarian states. Tegmark makes no predictions, but he does present many of the pros and cons of various futures. This is meant to spur consideration of the ethical and political future that we humans want so as to proceed with AI development and safety research that more adequately corresponds to our desires and best interests. He also discusses nuclear risks and other forms of existential destruction.

Chapter 6 speculates on life in the universe in the extremely far future. Tegmark engages such topics as making use of black holes for energy, colonizing cosmos, the death of the universe, and the complete rearrangement of matter by a superintelligence, and the unlikely (but potential) interactions between humans, AIs, and extraterrestrials.

Chapter 7 discusses the nature of goal-oriented behavior and design. Over time, with the generation of more complex forms of life and society, the universe appears more likely to be directed toward some kind of ultimate purpose. Tegmark discusses the evolution of the structure of goal-like behavior from basic physical reality to biology to psychology and beyond. Once we become self-conscious about goals, it is essential to set the most ethical goals we can in accordance with fundamental ethical principles. Tegmark also discusses the dialectical relationship between world-modeling and goal retention.

Chapter 8 discusses the nature of consciousness, which, for Tegmark, gives meaning and value to the universe. He identifies problems in consciousness research, ranging from scientific problems regarding how the brain works to fundamental philosophical inquiries about the nature of subjective experience and why subjective experience exists. Tegmark discusses various theories of consciousness and supports the idea that consciousness is an emergent property of our physical universe, thereby disagreeing with mind-body dualism. He returns to the concept of substrate independence and toys with the possibility of conscious AIs: “If artificial consciousness is possible,” Tegmark writes, “then the space of possible AI experiences is likely to be huge compared to what we humans can experience” (315).

The Epilogue outlines the birth and development of Tegmark’s nonprofit organization, The Future of Life Institute. Concerned primarily with the long-term future of intelligent life “through technological stewardship” (317) the Future of Life Institute, according to Tegmark, works to create consensus on the ethical development of AIs, amongst other things. Tegmark tells the story of a successful AI safety research conference and the large endowment provided to his organization by Elon Musk, founder of SpaceX and Tesla. Tegmark proposes an outlook of “mindful optimism” so that researchers and ordinary people can “think about what sort of future you want rather than merely what sort of future you fear” (334).