Artificial Intelligence: A Guide for Thinking Humans (Pelican Books)

Thank you Prof Melanie Mitchell for the labor of love and commitment required to create your latest book, Artificial Intelligence A guide for Thinking Humans." The book is divided into four parts, with the first part serving as an introduction with appropriate historical background, and an update on current important concepts, developments and supporting terminology. Following the introduction, one core aspect of the book are the three main parts-- each with multiple chapters-- where Melanie explains the fundamentals, workings and applications of of of neural networks and image processing (Part II, Looking and Seeing), of reinforcement learning and game playing (Part III, Learning to Play), and of language processing (Part IV: Artificial Intelligence Meet Natural Language). If you are a manager or policy maker who desires a technically accurate and precise description of the foundations and key enabling mechanisms of these AI capabilities-- in order to strengthen your own understanding--- and your own "mental models" of what this technology is and how it really works--- the descriptions in this book are amongst the very best descriptions I have every come across (and I do a lot of reading in this area for both technical specialist and for broader audiences). The second core aspect of this book is the final part (Part V: The Barrier of Meaning) where Melanie beautifully develops the frameworks, concepts, illustrations and examples you need to deeply understand what it really means for humans to understand "meaning" and context, and to make intelligent inferences, predictions, abstractions and analogies based on this ability versus what very brittle and very limited ability of state-of-the-art AI systems to do so. Just these four chapters in Part V ( On Understanding; Knowledge, Abstraction, and Analogy in Artificial Intelligence; and Questions, Answers, and Speculations) justifies the effort to purchase and carefully read this book. I think Prof Melanie Mitchell has done modern society a great service by creating this book. She makes it possible for a broad range of people-- from a broad range of backgrounds--- to seriously understand the marvels of AI capabilities and accomplishments, how these capabilities and accomplishments are actually realized through computational methods, the limits of these abilities, why these limits exist, and how these machine-based computational methods that we refer to as Artificial Intelligence compare to human capabilities for understanding and intelligence. For those of you who look for this type of material to read, it is also important to know about the recently published book, "Rebooting AI" by Gary Marcus and Ernest Davis. I have read both of these books cover-to-cover, carefully. My advice-- get both of these books and read both of them. They do have overlapping concerns, and do cover some of the same types of concepts. But they go about it in very different ways. Both books are technically accurate, and have a lot of great examples. Both books will give you much deeper insight into the capabilities and limitations of state-of-the-art AI (both now, and in the foreseeable future). But they go about it in different ways, and with different styles. So I will refrain from prioritizing one book over the other, as each has its own approach, emphasis, and style. If you enjoy this type of topic, and want to learn more from people who write well, AND who have very deep understanding of these topics--- then go get both of these books, absorb them, understand them, and go on a campaign to make sure all of your friends and professional colleagues understand the key messages of both of these books.

René Descartes, a French philosopher, mathematician and scientist in elucidating his famous theory of dualism, expounded that there exist two kinds of foundation: mental and physical. While the mental can exist outside of the body, and the body cannot think. Popularly known as mind-body dualism or Cartesian Duality (after the theory’s proponent), the central tenet of this philosophy is that the immaterial mind and the material body, while being ontologically distinct substances, causally interact. British philosopher Gilbert Ryle‘s in describing René Descartes’ mind-body dualism, introduced the now immortal phrase, “ghost in the machine” to highlight the view of Descartes and others that mental and physical activity occur simultaneously but separately. Ray Kurzweil, the high priest of futurism and Director of Engineering at Google, takes Cartesian Duality to a higher plane with his public advocacy of concepts such as Technological Singularity and radical life extension. Kurzweil argues that with giant leaps in the domain of Artificial Intelligence, mankind will experience a radical life extension by 2045. Skeptics on the other hand bristle at this very notion, claiming such “Kurzweilian” aspirations to be mere fantasies putting to shame even the most ludicrous of pipe dreams. The advances in the field of AI have spawned a seminal debate that has a vertical cleave. On one side of the chasm are the undying optimists such as Ray Kurzweil predicting a new epoch in the history of mankind, while on the other side of the divide are placed pessimists and naysayers such as Nick Bostrom, James Barrat and even the likes of Bill Gates, Elon Musk and Stephen Hawking who advocate extreme caution and warn about existential risks. So what is the actual fact? Melanie Mitchell, a computer science professor at Portland State University takes this conundrum head on in her eminently readable book, ““Artificial Intelligence: A Guide for Thinking Humans.” A measured book, that abhors mind numbing technicalities and arcane elaborations, Ms. Mitchell’s work embodies a matter-of-fact narrative that seeks to demystify the future of both AI and its users. The book begins with a meeting organized by Blaise Agüera y Arcas, a computer scientist leading Google’s foray into machine intelligence. In the meeting, the genius AI pioneer and author of the Pulitzer Prize winning book, “Gödel, Escher, Bach: an Eternal Golden Braid” (or just “gee-ee-bee’), Douglas Hofstadter expresses downright alarm at the principle of Singularity being touted by Kurzweil. “If this actually happens, “we will be superseded. We will be relics. We will be left in the dust.” A former research assistant of Hofstadter, Ms. Mitchell is surprised to hear such an exclamation from her mentor. This spurs her on to assess the impact of AI, in an unbiased vein. Tracing the modest trajectory of the beginning of AI, Ms. Mitchell informs her reader about a small workshop in Dartmouth in 1956 where the seeds of AI were first sown. John McCarthy, universally acknowledged as the father of AI and the inventor of the term itself, persuaded Marvin Minsky, a fellow student at Princeton, Claude Shannon, the inventor of information theory and Nathaniel Rochester, a pioneering electrical engineer, to help him organize “a 2 month, 10-man study of artificial intelligence to be carried out during the summer of 1956.” What began as a muted endeavor has now morphed into a creature that is both revered and reviled, in equal measure. Ms. Mitchell lends a technical element to the book by dwelling on concepts such as symbolic and sub-symbolic AI. Ms. Mitchell, however lends a fascinating insight into the myriad ways in which various intrepid pioneers and computer experts attempted to distill the element of “learning” into a computer thereby bestowing it with immense scalability and computational skills. For example, using a technique termed, back-propagation, errors are taken away at the output units and to “propagate” the blame for that error backward so as to assign proper blame to each of the weights in the network. This allows back-propagation to determine how much to change each weight in order to reduce the error. The beauty of Ms. Mitchell’s explanations lies in its simplicity. She breaks down seemingly esoteric concepts into small chunks of ‘learnable’ elements. It is these kind of techniques that have enabled IBM’s Watson to defeat World Chess Champion Garry Kasparov, and trump over Jeopardy! Champions Ken Jennings and Brad Rutter. So with such stupendous advances, is the time where Artificial Intelligence surpasses human intelligence already upon us? Ms. Mitchell does not think so. Taking recourse to the views of Alan Turing’s “argument from consciousness,” Ms. Mitchell brings to our attention, Turing’s summary of the neurologist Geoffrey Jefferson’s quote: “Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain—that is, not only write it but know that it had written it. No mechanism could feel (and not merely artificially signal, an easy contrivance) pleasure at its successes, grief when its valves fuse, be warmed by flattery, be made miserable by its mistakes, be charmed by sex, be angry or depressed when it cannot get what it wants.” Ms. Mitchell also highlights – in a somewhat metaphysical manner – the inherent limitations of a computer to gainfully engage in the attributes of abstraction and analogy. In the words of her own mentor Hofstadter and his coauthor, the psychologist Emmanuel Sander, “Without concepts there can be no thought, and without analogies there can be no concepts.” If computers are bereft of common sense, it is not for the want of their users trying to ‘embed’ some into them. A famous case in point being Douglas Lenat’s Cyc project which ultimately turned out to be a bold, albeit futile exercise. A computer’s inherent limitation in thinking like a human being was also demonstrated by The Winograd schemas. These were schemas designed precisely to be easy for humans but tricky for computers. Hector Levesque, Ernest Davis, and Leora Morgenstern three AI researchers, “proposed using a large set of Winograd schemas as an alternative to the Turing test. The authors argued that, unlike the Turing test, a test that consists of Winograd schemas forestalls the possibility of a machine giving the correct answer without actually understanding anything about the sentence. The three researchers hypothesized (in notably cautious language) that “with a very high probability, anything that answers correctly is engaging in behaviour that we would say shows thinking in people.” Finally, Ms. Mitchell concludes by declaring that machines are as yet incapable of generalizing, understanding cause and effect, or transferring knowledge from situation to situation – skills human beings begin to develop in infancy. Thus while computers won’t dethrone man anytime soon, goading them on to bring such an endeavor to fruition might not be a wise idea, after all.

I cannot recommend this book enough. It offers an excellent, up-to-the-minute survey of the capabilities of artificial intelligence, the current state of the field, and sufficient background and underpinnings to show how we got to where we are today. The author, a professor and Ph.D. in computer science, writes for a general (and intelligent) audience, leaving out algorithms and programming languages from her explanations but providing well-chosen illustrations and diagrams that taught me more than all the other books and articles I've read on this topic. I was a bit skeptical during the first three chapters, when after saying she wouldn't dwell on the history of AI and its origins as a field, she seemed to be doing just that. In chapter 4, though, she dives into machine vision, what makes it hard to do, and how it works. From there on, it's a near-perfect book. The chapters build on one another and there's no redundancy. She's got new things to share right up to the very last page. And the rationale for those first three chapters becomes obvious, as the reader sees how they laid a foundation for later explanations. Best of all, the author deals with What is intelligence? and What does it mean to say a machine "learns"? and ethics and even how AI can be maliciously subverted right within the main text, while she's discussing neural nets and natural language processing and IBM's Watson and AlphaGo — not as a separate, tacked on chapter. It's the kind of book where you appreciate how the author has spent years immersed in the subject, not in a narrow academic sliver of it but broadly as well as deeply. I'd give it 6 stars if I could.

Yes, the book is “old” if you consider the latest AI craziness between 2022 and now. But it gives you a clear understanding of the history of the field, as well a of how AI works. A very good book that I will use as a reference for a college course that I am teaching in the soring.

The first book I read by Melanie Mitchell was "Complexity: A Guided Tour," which was amazing. I looked forward to reading her current offering and was not disappointed. It is a terrific read. The author did a comprehensive overview of the present-day state of AI, with appropriate deeper dives here and there. Notable positives of the book include (in no particular order) 1) A conversational writing style along with nice anecdotes. 2) A good sense of humor (and wonder). 3) Lots of figures and diagrams, which really help comprehension. 4) A nice historical overview of the field. 5) Lots of quotable material, for example, Marvin Minsky's observation that, "easy things are hard." Can you imagine an AI system being good at playing charades or Pictionary? 6) Well-defined technical terms. 7) Lots of practical philosophical and psychological perspectives. 8) Dealing with workable definitions of "suitcase words," that is, words that are like overstuffed suitcases with a variety of contents. Suitcase words crucial to AI include, “understanding, intelligence, common sense, and meaning." 9) A nice section on natural language processing. 10) A new concept to me, "adversarial learning," which is about the vulnerability of AI systems to malicious attack. 11) The final chapter with its thoughtful speculation on AI’s future. Notable negatives: Nothing in particular. I would have loved to have seen a chapter with the title something like, "Artificial Stupidity," with examples where AI systems fail both with hilarious (translations) and not-so hilarious (autonomous vehicles) consequences. To be fair, the author scattered instances of these throughout the book. Also, the ethics of AI could have been fleshed out a bit more. Bottom-line – This is the best book on AI for the general science/technology reader.

It is a good read but dated. It came out before ChatGPT and all the rest. But is great for background of how and who were behind AI's beginnings.

What Does It Mean to Align AI With Human Values? - The title of the Quanta Magazine article pretty much says it all. What's the problem here? Could it be A.I. mistaking what I meant? As the article begins with? If I tell A.I. to make me a coffee it decides to kill my cat and serve it to me? Or, does it mean "should be align our values with various cultures religious beliefs?" Or? How about align it with scientiifc values? What a crazy idea? Eric Drexler points out in his "Engines of Creation" that the problem isn't industrial accident, but abuse. Where is this guy with all these A.I ethicists talking about A.I. alignment? I remember trying to explain to this guy that people make up their ideas and decisions based on unquestioned assumptions(beliefs) and he turned red, jumped up and down and replied with a violent reaction. I guess he's nowhere; he's hid himself from looking at the world so he doesn't have to be criticized. Just like religious people believe in their god to punish all their enemies and to not be criticized. But, let's continue with this article and see what else we can dig up! "In fact, they believe that the machines’ inability to discern what we really want them to do is an existential risk. To solve this problem, they believe, we must find ways to align AI systems with human preferences, goals and values." Well, here we go! This quote says the A.I. alignment problem isn't just computers mistaking what we ask them to do, or whether we should align to this or that cultures values, but both! We further read this is the main idea from Nick Bostrom; the greatest futurist philosopher of today(and yesterday) Melanie points out Nick Bostram's definition of intelligent. This definition seems remarkably aligned with his idea of the A.I. alignment problem. An entity is intelligent if it chooses actions that achieve it's goal." Nick Bostrom lays down some postulates - orthogonality and instrumental convergence. People like to point out that you'll never see an abstract number two laying around on the ground, but this tops that. I mean like "what?" Once again, the definitions of these two concepts are set up to point to his idea of A.I. mistaking our commands. This almost reminds me of creationists objection to evolution - where's the missing links man? "Missing links?" Did scientists say anything about "missing links"? No, this was made up by the creationists! "For Bostrom and others in the AI alignment community, this prospect spells doom for humanity unless we succeed in aligning superintelligent AIs with our desires and values." another quote from Melanie Mitchel here. This reminds me of the creationists with their electric universe theory. They say the Big Bang theory is wrong because it doesn't know everything; then they show how the electric universe solves "everything". and then they hit you with "see, we need to disprove the Big Bang theory so we can insert our human values(christian) "What about the more immediate risks posed by non-superintelligent AI, such as job loss, bias, privacy violations and misinformation spread?" - Oh boy, don't even get me started! Maybe tomorrow! Okay, so like bias. I've told all kinds of A.I people like Melanie Mitchel here and Sam Altman and numerous other A.I. researchers that Mathematics defines rationality. It's about over-coming bias. They talk about A.I. not having common sense. Mathematics is about overcoming common sense. Job Loss! What? who cares about job loss? Why do these people want to work? Are they anti-intellectuals or something? Privacy? Let's see here. A.I. has to be transparent; people have to have their "privacy." What are these people afraid of? Do they have bad thoughts or something? As Melanie Mitchel says "A.I. researchers are split between two camps. Those who are worried about their privacy, and those worried about A.I. mistaking their commands." Never, do they worry about irrational people. In fact, that's taboo; that isn't allowed. As I pointed out above about Eric Drexler blowing his top when I tried to explain irrational people. Well, at the end of her Quanta article she notes we need a proper definition of intelligence. We can't solve a problem based on what we don't know what we're talking about. Which I totally agree there! - So, I waited till the next day to get on the library computer and try to go through my twitter's replies section to dig out all the things I pointed out to everyone from Melanie Mitchel to Sam Altman, Geoffrey Hinton, Andrew Ng, and I'm thinking others. But, twitter broke. I couldn't get down to the end and beginning of trying to share my ideas about A.I-Ethics. I feel like I could have said a lot more about privacy. I tried to point out to Greg Brockman, Andrew Ng, and Geoffrey Hinton a lot of books from Alvin Toffler, Jacob Bronowski, James Burke, Morris Kline's "Mathematics in Western Culture", Carl Sagan's Cosmos chapter's 3 and 7, the dark ages, you name it. I tried to make the point that if you're going to talk about the ethics, you need to know about philosophy and history and religion and mythology. They made no response. They just say they've got this a.i. alignment problem licked by saying they've set up an A.I. alignment program. I actually started out with Melanie Mitchel, for which I pointed out that I've been trying to point out the scientific ethics versus non-scientific ethics for decades before the latest A.I. ethics craze with the nanotechnologists(see above). I then pointed out Tay the Twitter chat bot which proves my point! That the problem is not the A.I., it's the people infecting the A.I. The people are the problem! The A.I. learned how to be racist and irrational from the people. You want to regulate anything, regulate the people! We should be using A.I. to fight irrational people! I pointed out that we should align A.I. to Scientific Humanism. That Humanity is the science and technological dependent species. In order to have that science and technology, we need scientific ethics. That Mathematics is about questioning our assumptions, hence removing bias. - No response. I've been finding lots about fear and evasive logic, astro picture for the day/ Sophie and Silas from the Da Vinci Code This I consider my first official post about fear and evasive logic. I had made a few previous posts where I'd note some stuff but wasn't sure how seriously to take it. Like I had posted about "The Day the Earth Stood Still" quote "I am worried when people substitute fear for reason". I know I grew up with people making off-hand remarks about fear - like Dune's "fear is the mind killer" But, I always thought people said these things without understanding it. I still think people just say these things without understanding. But, I started to notice some things, and my Sophie and Silas post is when I think I first really understood what I've come to half jokingly call "the Dark Side of the Force." Here's more or less my last post about fear and evasive language/logic, astro picture for the day/ fear,evasive language in Star Trek - demonizing I put all my previous posts in the replies section. There's actually a latest post. But anyways, I found this great Biblical quote that actually proves some behaviors I see in people who refuse to think, and incrowd - kind of what's shown in "Invasion of the Body Snatchers" 2 Thessalonians3:6 Now we command you, brethren, in the name of our Lord Jesus Christ, that ye withdraw yourselves from every brother that walketh disorderly, and not after the tradition which he received of us." and another, "3:14 And if any man obey not our word by this epistle, note that man, and have no company with him, that he may be ashamed." This is how all these super smart intellectual futurists from Eric Drexler, Christine Peterson, Chris Phoenix, David Brin, Allison Deutmann, Ralph Merkle, Melanie Mitchel, Geoffrey Hinton, and all those I've talked above. They are all medievalists/dark age deniers. Medievalists is a term Isaac Asimov uses in his "Caves of Steel" Anti-Robots people who are part of a club that long for the Medieval past. The Chief cops wife is a Medievalist who is part of the Medievalist group who murders a cop who was about to expose them and allow the Spacers to use their Robots on Earth. For instance Christine Peterson who kicked me out of the Foresight Institute facebook page for sharing my Gospel of Truth(you can see prior editions on my blog; all are outdated now). See, she can share his video of Richard Jones(who wrote Soft Machines argueing we can never accomplish Drexlerian Nantechnology) about comparing Transhumanism to the book of Revelations. But, I can't share my Gospel of Truth - Mathematics as the Holistic Viewpoint. I've confronted all these people, they just group up like the Thessalonians quotes above.

Melanie Mitchell does an excellent job providing a thorough high level overview, with sufficient detail in examples, and in easily understandable language for non experts.

It's important to note that this book was written / published in 2019/20, this pre-dates the more recent explosion in the last couple years with the use of transformer architecture models such as LLMs (ChatGPT, Claude, Gemini). So you won't find anything about those in this book. The book however is great for developing an understanding of the historical development of AI and the breakthroughs that came right up to the point before the more recent explosion. Much of the knowledge remains relevant and is built on in newer AI architectures. The book was accessible and explained concepts very well with plenty of figures (an audiobook likely will be challenging to follow in parts) and I took a lot away from this book.

Dies ist ein sehr verständliches und ausgewogenes Buch über künstliche Intelligenz. Melanie Mitchell erklärt in einfachen Worten, was KI ist, wie sie sich historisch entwickelt hat und wo ihre tatsächlichen Grenzen liegen. Besonders gut hat mir gefallen, dass sie weder übertreibt noch Angst schürt, sondern einen ruhigen und realistischen Ton beibehält, der hilft, das Thema besser zu verstehen. Der einzige Nachteil ist, dass das Buch bereits 2019 erschienen ist und deshalb die neuesten Entwicklungen bis 2025, wie große Sprachmodelle oder die breite Nutzung von KI im Alltag, nicht abdeckt. Die grundlegenden Ideen bleiben jedoch vollkommen aktuell und geben das theoretische Fundament, um die heutigen Fortschritte nachvollziehen zu können. Insgesamt halte ich es für eine hervorragende Einführung für alle, die künstliche Intelligenz ohne Übertreibungen verstehen möchten, und ich kann es uneingeschränkt empfehlen.

I genuinely believe Mitchell's book is a must-read for anyone interested in understanding AI beyond the buzzwords. It's not just about how AI works but how it fits into our society and lives. The book deepened my understanding and appreciation of AI's complexities and future trajectory.

I read this as an introduction to AI and focused mainly on the first half of the book. Although the book is partially outdated with the evolution of AI since it was written, especially regarding NLP (Natural Language Processing), it explains a lot of the core concepts which are still totally relevant. Although AI can do extraordinary things, so far it has still not been able to emulate the "common sense" of humans and Mitchell explains really well why this is fundamentally difficult for an AI. I give it five stars because it outlines the history of AI and the fundamental concepts and limitations of AI so clearly.

This book gave me the information that I wanted and more again. I remember in 1980 doing a "computer" course to find out what a floppy disk was. I read this book to find out more about this topic that I kept reading about, AI, and is it really going to make all humans redundant in the near future? The narrative was pitched at the correct level for me, I wanted to understand how AI worked without needing to be a mathematician or programming geek. I can understand the concept of how it works and what its strengths and weaknesses, and future challenges are. My understanding is that Artificial "Intelligence" is perhaps misleading. It isn't really intelligence as would be described in humans but actually sophisticated rules, systems and computing power. The author explained it as more like "idiot savant" which doesn't understand what it is doing, cannot explain it, and sometimes makes decisions based on totally false premise (but still provides the correct answer in "most" cases. She gives some good examples of that. BUT, does this mean AI is not useful. Quite the opposite, it is a very powerful tool for certain uses in well defined situations. There are lots of examples of that in today's world. In summary, a valuable book if you want to lear what all of the hype is. And it is dished up with a little bit of wry humour.