Curated by UCL

User Reviews

  • 1 June 2016 at 6:59 pm

    What to Think About Machines That Think: Today’s Thinkers on the Age of Machine Intelligence is a collection of short essays that discuss artificial intelligence (AI) and the possible implications of this set of technologies on humanity. The book shines a light on the (impressive, and in some cases even unnerving) capabilities that intelligent computer systems have recently acquired as a result of a number of incredible advances, many of which have gone relatively unnoticed outside of the scientific community. In an effort to wrest the details from this small and exclusive group of specialised computer scientists, the editor John Brockman features a diverse cast of 170+ authorsincluding business people, artists, journalists and academics of all disciplinesto ultimately set the tone for a new and very serious public conversation about where this technology could, and should, lead to.

    AI, a classic discipline of computer science, is concerned with reproducing intelligent behaviour in computers. Its modern history dates back to the 1950s and is infused with cycles of sky-high expectations and a subsequent inability to live up to the hype. This has lead to those in the field being rather careful about labeling any progress as `groundbreaking’. In recent years, however, research has given rise to a number of extraordinary developments, leading many to speak of a true renaissance in AI research.

    While it would be a major overstatement to announce the advent of human-like intelligence (the discipline’s holy grail), it is expected that AI, and particularly machine learning, will soon enter our lives at a much deeper level than ever before. The full potential of current-day machine learning is mostly embodied in business tools, rather than consumer products. As a result of this, the general public has yet to fully partake in this conversation. By asking leading thinkers for their thoughts, hopes and fears regarding AI, John Brockman assembles the conflicts and questions that define the field today and, in doing so, opens the discussion up for everybody.

    Among the essays (very few of which stretch longer than five pages), it is remarkable how the authorsindividuals with a deep knowledge of the same subject matterdiffer profoundly in terms of what makes them worried or excited about AI. The approaches taken range from machine consciousness, and whether or not we should fear apocalyptic world-domination scenarios (spoiler: no need to panic), to socially structuring a world in which machines do all of the work. Here it should be noted that while it is understandable that every author wishes to reinforce why they feel that AI is highly impactful, it makes for an unfortunate element of repetition when reading multiple essays in one go. This is just a tradeoff that comes with the format of loosely chaining together self-contained essays (each of which requires a separate introductory paragraph) and is more than made up for by the wide thematic diversity among them (after the repetitive introductory paragraph).

    Readers looking for hard facts, technical details and an idea of `how stuff works’ should probably stay away from this book. The authors tend to fast-forward over the creation of intelligent systems, and set off with the premise that strong general AI has already been built. Although this is certainly what the volume set out to cover, it prompts some writers to treat AI as an abstract or even mythical force. Considering how deeply involved some of them are with cutting-edge AI research, the lack of insider perspectives on the actual technology at times induces an urge to grab and shake the author to make them reveal their tech secrets.

    Overall, the result of this highly philosophical and moralistic approach is double-edged. Some writers shine as a result of the freedom and build truly thought-provoking arguments. For example, the theoretical physicist Antony Lisi notes that AI systems would not be the first non-human autonomous entities to take on an influential role in human societies – he argues that corporations enjoy the legal rights of humans every day. In another essay, computer science professor Jon Kleinberg argues that the widespread fears of conscious machines that emotionally compete with humanity are a product of science-fiction culture. He claims that people should be much more troubled by the fact that algorithms, which influence millions of lives everyday, are reaching levels of complexity that are simply incomprehensible for human minds. And if we cannot understand the data-based decisions that computers are making for us, are we really still in control?

    In contrast, the book also features several pieces that fail to make it past the standard “It will change the world – for better or worse” attitude, which makes starting a new essay something of a lottery. Coupled with the editor’s decision not to order the essays using any apparent convention, this makes organised reading of this book harder than it should be.

    In conclusion, What to Think About Machines That Think is an accessible gateway for readers with an interest in AI. The book manages to offer compelling alternatives to the typical black and white opinions regarding the field, and does its job in widening horizons. However, don’t expect this book to make you an expert in AI. Ultimately, it is not a science book, but rather a book about society and how it can deal with technology.

Add a Review

To leave a review you need to be logged in to your account on ENGins.

Sign up for free or log in if you already have an account.