In Our Own Image: Savior or Destroyer? The History and Future of Artificial Intelligence

by George Zarkadakis

A timely and important book that explores the history and future, as well as the societal and ethical implications, of Artificial Intelligence as we approach the cusp of a fourth industrial revolution Zarkadakis explores one of humankind's oldest love-hate relationships―our ties with artificial intelligence, or AI. He traces AI's origins in ancient myth, through literary classics like Frankenstein, to today's sci-fi blockbusters, arguing that a fascination with AI is hardwired into the human psyche. He explains AI's history, technology, and potential; its manifestations in intelligent machines; its connections to neurology and consciousness, as well as―perhaps most tellingly―what AI reveals about us as human beings. In Our Own Image argues that we are on the brink of a fourth industrial revolution―poised to enter the age of Artificial Intelligence as science fiction becomes science fact. Ultimately, Zarkadakis observes, the fate of AI has profound implications for the future of science and humanity itself.

Genres: Nonfiction, Science, Artificial Intelligence, Technology, Philosophy, History, Audiobook

384 pages, Hardcover

First published March 5, 2015

Reviews

[Alicea · Rating 4 out of 5]

I started reading In Our Own Image: Savior or Destroyer? The History and Future of Artificial Intelligence by George Zarkadakis with the naive idea that he would definitively answer the question "Will robots destroy the human race?" Instead he's done something even better. He's given a comprehensive history of mathematics, philosophy, psychology, computer science, and Artificial Intelligence. He's explained how our evolution as a species is directly related to the possibility (and the success) of the creation of AI. I'm always amazed when reading scientific nonfiction to discover just how much of the history I already know (Turing! Aristotle!) and how much I am woefully ignorant of (Godel and so much more). AI has been in the works for years. Scientists have approached it from a variety of standpoints and for a multitude of purposes. One of the biggest hurdles facing the scientific community is whether or not consciousness is programmable in a computer. A machine can be "intelligent" but does it exhibit intelligence in the same way as humans? Can a machine understand the true meaning of the questions that it answers through the logic programs installed in its hardware? No one really knows. Is it possible to map the human brain and recreate this mechanically inside of an android? Probably. Is humanity ready for machines that look like us, act like us (to an extent), but are more efficient and intelligent than us? Doubtful. As you know, I love a good end notes and this book has a truly excellent one. I learned the difference between ontology and epistemology, that there's something called the Ig Nobel Prize, and that writers are simply "empty vessels" waiting to be filled by texts. This book asks more questions than it answers and I think that's the point. Humans are unique because we have the capacity to ask questions and to be curious. If you're looking for a thought provoking book on a truly fascinating (and still terrifying) topic then this one will definitely fit the bill.

[Andrea · Rating 5 out of 5]

A book right up my alley: a perfect marriage of technology and philosophy with a generous dash of history. It touches upon our fascination with replication of life, that stretches all the away back to 40,000 ago; summarizes some of the greatest thinkers' theories on mind and consciousness, and how it would apply to thinking machines; and looks ahead at what might lie ahead for us in pursuit of AI. There was so much I wanted to annotate, so I am going to purchase my own copy and read it once again at some point in the near future. Definitely a keeper for a personal library.

[Ihor Kolesnyk · Rating 4 out of 5]

Історія ідей дуже цікава. Метафори також. Загалом книга добра, саме тому 4/5

[Henry Manampiring · Rating 4 out of 5]

I like it. Sometimes it can get dizzying with technicalities and terminologies, but overall it is enlightening on AI subject. Note: this was written before ChatGPT rise to fame.

[Natali · Rating 5 out of 5]

An absolute masterpiece. I wish I hadn't listened to the audiobook. I have to go back and buy a written copy because it is so dense.

Zarkadakis presents artificial intelligence not as a sudden phenomenon, but as the culmination of centuries of myth, philosophy, and storytelling. He weaves together everything from ancient fables to Enlightenment thought to modern tech, showing how we’ve been slowly — and perhaps inevitably — marching toward this moment.

I was afraid of AI before this book but now I feel more than ever that we have to take a collective hand at shaping it. I will refer to this book again and again as we move into this monumental revolution!

[Rhys · Rating 2 out of 5]

When discussing computers and AI it was very good. The discussion on the philosophy of mind was pretty good. The emergence of autocatalytic systems was interesting. The reaches into political philosophy and economics were, well, reaches.

Overall, a raucous read with some gems.

[Pat Rolston · Rating 4 out of 5]

There is much more history rendered in this book regarding the development of computing, logic, philosophy, and human anatomy versus AI as implied by the title. That is by design as all these disciplines are elements of AI allowing for a better grasp of the challenges obtaining functional artificial intelligence and what that implies. He delves deeply into the various interpretations of consciousness as the ultimate holy grail. This in theory is the ‘ghost in the machine’ once global AI’s potential is fully realized.

I felt the author did a very credible job with this approach while falling a bit short once he focused on the final stretch and actual AI and his conclusions. He does an expert job taking such vast and complex topics yet distilling them in an effective, entertaining manner. Highly recommended for anyone wanting to delve into what may be the most important subject mankind faces for generations to come.

[Kat · Rating 4 out of 5]

I won this book as a Goodreads giveaway.

This book really got me thinking about the way society views Artificial Intelligence and the technological advances that are being made to bring us closer to this idea. It states that people have a fascination with AI because of an early influence through art or movies, maybe even just the idea of creating something that can do and say whatever we want it to. But the one thing we want most in an AI is for it to "feel" or "think" the way we do. To create something that is like us that can have emotion and depth like a human. I found this the most interesting concept in the book because the author states that for something to have these qualities it must have conscience. I don't necessarily have the same beliefs as the author but this got me thinking on a higher more spiritual level.

Overall this book was informative but imaginative.

[William · Rating 5 out of 5]

Great book that mixes philosophy, quantum mechanics, history, pop culture analysis and speaks on every field with authority and style. I came away thinking the author is not particularly convinced that AI is possible, but that perhaps AI will take a different form than our idea of consciousness. But the analysis isn't necessarily "anti-AI," just professing a healthy skepticism that the idea of machine consciousness may be much more complicated than we believe.

[Lukas Vermeer · Rating 4 out of 5]

Insightful. Goes deep on context (instead of content) of current AI hype.

[Ana · Rating 1 out of 5]

Very conflicted read. I was really excited by the premise of the book and the author, by their setting out to figure out our fascination with artificial intelligence/life by looking at the history of these concepts in literature, philosophy etc.
But it did not deliver. I found some takes to be rather naive and oversimplifying, for example: "Loneliness and the inability to feel love usually lead to depression, suicide even." - this sentence destroyed me a little (maybe I am too sensitive), but there is an abysmal lack of depth (pun intended I guess?), a shallow oversimplification that illustrates almost nothing but shrouds so much.
I'm not really an expert in anything, but even so the writing felt very dilettantish - like sitting with a friend in a café and them rambling about some interesting idea they read about: nice and intriguing enough for a coffee chat, but somehow I expected a bit more depth of analysis when the ideas are then gathered together in a book.
About half-way through the book it got worse, when thermodynamics came into the discussion. It was claimed (maybe it is just an editing issue?) that "[...] say we heat up a container filled with gas - gas molecules start speeding around unpredictably. The order to the system increases, and entropy decreases." - unless there are some other assumptions about the system which were not mentioned, this is just wrong. In such a situation order decreases, disorder goes up and so does entropy. But I know first-hand how challenging wrapping one's head around thermodynamics is, a never-ending quest really, so I forgave this slip. However, a few pages later it is claimed that using EEG (electroencephalography) and MEG (magnetoencephalograohy) in conjunction with fMRI (functional magnetic resonance imaging) "researchers can correlate what happens at the level of a single neuron". (!!!) But these methods on their own only measure average activities of groups of neurons, they do not have neuron-level resolution, and as such the statement should be either changed or there should be an explanation regarding how the neuron-level resolution can be achieved (inference models? data coming from intracranial electrodes?). I know very little of these methods, but enough to feel like that statement is quite misleading, especially for someone completely new to the topic. I lost the little faith I had at this point, and skimmed-read the second half of the book.
What also depresses me about this is the quote on the cover of the book: "Timely and important" by Jim Al-Khalili. I already have "Life on the edge" by Jim Al-Khalili on my shelf, but I am now already wary of whatever claims might be made there as well.

I can't decide on the ratio between this book being just badly written vs actually wrong. I wrote this review because if someone else had written something like this I probably wouldn't have picked this book up, which I wish I hadn't.

(later edit a few hours later): The positives include the philosophical bits, the duality of mind vs matter discussions, discussing how empiricism enabled science to shed light on so much of our reality and the like. But even though I enjoyed these parts, I did so with a grain of salt because I don’t have a background in philosophy, so I can’t really tell if something is off or not, like I was able to for other parts of the book. (or maybe I expect too much thoroughness/rigour from a popular science book and need to chill).

[Ramon van Dam · Rating 4 out of 5]

Not quite what I expected, a very interesting read. I thought this would mainly be a book about the current state of AI with a focus on tech, but I was pleasantly surprised to find out that Zarkadakis spends a huge portion of the text to discuss more philosophical questions.

What is consciousness, what does it mean to be alive, what has humanity already tried to simulate it and are they realistic approximations? The author tries to answer these questions and more in a very pleasant way. A purely computer science focused book would also have been interesting for me personally, but the angle that he took definitely bumped it up a notch.

[Alexandra · Rating 4 out of 5]

A satisfactory book on an intriguing subject, although I can't resist pointing out a particular recurring error: apparently, there's something called the "Plank length".

[BLACK CAT · Rating 5 out of 5]

A tour of our first attempts at artificial intelligence from the ancient Greeks to the future possibilities with neuromorphic engineering and autopoiesis.

[A A · Rating 3 out of 5]

At first glance, Zarkadakis's book is a history of artificial intelligence.

But underneath, it is really a history of human self-understanding.

His story moves through three great questions:

Why have humans always wanted to create artificial life?

What exactly is the human mind?

Can we reproduce intelligence—and perhaps eventually consciousness—in machines?

The remarkable insight is that these are not three separate questions.

They form a single chain:

What we build → reflects what we believe we are → changes how we understand ourselves.

PART I — DREAMING THE ELECTRIC SHEEP
Humanity's ancient dream of artificial life

The story begins long before computers.

Human beings have always imagined creating something in their own image.

Ancient myths contain artificial beings, animated statues, mechanical servants and artificial creatures. Greek stories of automata already imagined objects that could move and perform tasks.

Why is this important?

Because it reveals a very old human desire:

If nature can create intelligent life, perhaps human beings can create it too.

The dream is not merely technological.

It is almost existential.

We don't simply want machines that work.

We want to know whether life and intelligence can be recreated.

From gods to engineers

There is a gradual movement from:

magic → mechanism → science → computation → AI.

Ancient humans imagined gods creating life.

Later humans imagined machines creating lifelike behavior.

Modern humans imagine algorithms creating intelligence.

The technology changes, but the underlying dream remains remarkably similar:

Can the created become like the creator?

That question will eventually become central to AI.

PART II — THE MIND PROBLEM
What exactly are we trying to recreate?

Here Zarkadakis turns from artificial beings to the human mind itself.

And suddenly the problem becomes much harder.

We know how to describe some things about the brain.

We know about:

neurons,
electrical signals,
sensory processing,
memory,
learning,
brain regions,
neural networks.

But we still face the enormous mystery:

How does physical matter produce subjective experience?

Why does seeing red feel like something?

Why does pain hurt?

Why does music move us?

Why is there an inner point of view?

This is the famous mind–body problem.

Descartes: the great separation

René Descartes gave Western philosophy one of its most influential frameworks:

body = physical matter

mind = thinking substance

This is Cartesian dualism.

It gives us an intuitive explanation for consciousness:

Perhaps the physical brain is not the whole story.

But it creates another problem:

If mind and body are different substances, how do they interact?

How does an immaterial thought cause a physical arm to move?

That question remains unresolved.

Hobbes and the mechanical human

Thomas Hobbes pushed toward the opposite conclusion.

Perhaps thinking itself is physical.

Perhaps:

sensation → memory → imagination → reasoning

can ultimately be explained through mechanisms.

This opens the door to a radically different possibility:

Perhaps intelligence does not require anything mysterious.

Perhaps it is something a sufficiently complicated physical system can do.

That idea becomes enormously important for AI.

The cybernetic revolution

The twentieth century introduces another transformative idea:

Cybernetics

Cybernetics studies control, communication and feedback in systems.

Consider a thermostat:

sense temperature → compare → act → sense again → adjust

The system continuously uses information about its environment to regulate itself.

Scientists began seeing similar patterns in living organisms.

The brain became increasingly understandable as:

a system that receives information, processes it, acts on the world and learns from feedback.

This creates the idea of the cybernetic brain.

And that is a turning point.

The brain is no longer imagined simply as the mysterious seat of the soul.

It begins to look like an information-processing system.

The computational theory of mind

This leads to one of the most influential modern ideas:

Perhaps the mind is essentially computation.

The analogy becomes:

brain → hardware

mind → software

If that is correct, then something astonishing follows:

The same computation might theoretically be implemented in different physical materials.

Perhaps:

neurons can compute

and therefore:

machines might compute too.

This becomes one of the philosophical foundations of artificial intelligence.

But computation creates a new problem

Suppose a machine can:

speak,
reason,
recognize faces,
remember,
solve problems,
write poetry,
imitate emotion.

Does that mean it experiences anything?

Not necessarily.

And this distinction becomes fundamental:

Intelligence

Ability to solve problems, learn, predict and act.

Consciousness

Having subjective experience.

A machine might have the first without the second.

The philosophical zombie

Imagine a creature that behaves exactly like a human.

It says:

"I'm in pain."

It cries.

It avoids fire.

It describes its emotions.

But suppose there is nothing being experienced internally.

There is no subjective feeling.

This thought experiment challenges the idea that outward behavior is enough to establish consciousness.

And that becomes crucial for AI.

A machine might behave exactly like a conscious being while we still have no definitive way to know whether it actually experiences anything.

Searle's Chinese Room

John Searle gives another challenge.

Imagine someone who doesn't understand Chinese sitting inside a room.

Chinese symbols enter.

The person follows a giant rulebook and sends appropriate Chinese symbols out.

From outside, the person appears to understand Chinese.

But internally, they understand nothing.

Searle's argument:

Manipulating symbols according to rules may not be the same as understanding their meaning.

In its simplest form:

syntax ≠ semantics

A computer can manipulate symbols.

But does it understand them?

That question remains central to AI.

Connectionism

Another theory takes a different route.

Instead of treating the mind as a traditional symbolic computer, connectionism sees intelligence emerging from networks of interconnected units.

This resembles biological neural networks.

Rather than explicitly programming:

"If X, then Y."

we allow a system to learn patterns from examples.

That idea eventually leads toward modern neural networks and deep learning.

Embodied cognition

Then another challenge appears:

Maybe intelligence isn't just inside the brain.

Perhaps it emerges from:

brain + body + environment.

Think about riding a bicycle.

You don't learn merely by memorizing rules.

You learn through:

movement,
balance,
vision,
touch,
gravity,
feedback,
mistakes.

This suggests that intelligence may be embodied.

A disembodied computer might therefore not possess the same kind of intelligence as a human being.

Extended mind

And then the boundary expands again.

Perhaps cognition can extend beyond the brain.

A notebook can store memory.

A calculator can perform calculations.

A smartphone can provide navigation.

An external tool can become part of our cognitive process.

This gives us the extended mind idea:

The mind may not end neatly at the boundary of the skull.

Human intelligence is therefore increasingly understood as something distributed across:

brain + body + tools + social environment + culture.

PART III — ADA IN WONDERLAND
From logic to artificial intelligence

Now Zarkadakis asks:

If intelligence can be described formally, can we actually build it?

And the story goes back to logic.

Aristotle: can reasoning be formalized?

Aristotle developed formal logic.

For example:

All humans are mortal.

Socrates is human.

Therefore, Socrates is mortal.

The crucial discovery is that reasoning has structure.

If the structure can be formalized, perhaps it can eventually be mechanized.

This is the beginning of a long intellectual road toward computing.

The liar paradox

Then comes:

"All Cretans are liars."

If the speaker is telling the truth, he is lying.

If he is lying, perhaps he is telling the truth.

This apparently simple paradox exposes problems within formal reasoning.

And centuries later, such problems lead mathematicians toward deeper questions about the limits of formal systems.

Gödel: mathematics has limits

Kurt Gödel demonstrated something extraordinary.

Very roughly:

A sufficiently powerful formal system contains truths that cannot be proven within the system itself.

This challenged the dream of mathematics as a completely closed logical machine.

And it raises a provocative question for AI:

If human mathematical insight sometimes escapes formal proof systems, can human intelligence really be reduced completely to computation?

There is no simple answer.

Babbage and Ada Lovelace

Charles Babbage imagined the Analytical Engine.

Ada Lovelace saw something even more profound:

A programmable machine might manipulate not merely numbers but representations.

Numbers could represent music, symbols or other forms of information.

That is a crucial step toward the modern computer.

But Lovelace also raised an objection:

Machines execute what humans know how to instruct them to do.

This becomes the Lovelace objection.

It anticipates today's question:

Does AI genuinely create, or does it merely produce consequences of human-designed procedures?

Turing

Alan Turing then changes the question.

Instead of asking:

Can machines think?

he asks us to consider whether a machine can behave in ways indistinguishable from human intelligence.

This produces the famous Turing Test.

It represents a philosophical shift:

from asking what something is internally

to

asking what it can do.

That distinction remains enormously important today.

Bletchley Park

During World War II, Turing and his colleagues used computation for cryptanalysis.

This was a decisive moment.

Computation was no longer simply theoretical.

Machines were performing intellectual operations with enormous speed.

The dream of mechanical intelligence suddenly looked much more realistic.

Shannon and information

Claude Shannon gives the digital world another foundation:

information can be represented mathematically as bits.

Ultimately:

text → bits

sound → bits

images → bits

instructions → bits

This means enormously different forms of human information can be processed by the same general computational machinery.

The world becomes increasingly digitizable.

Von Neumann and stored programs

Modern computers become general-purpose because instructions themselves can be stored in memory.

The machine doesn't merely process data.

It can process programs.

This produces the modern distinction:

hardware

versus

software.

And again, the analogy with the human being appears:

brain : mind

as

hardware : software

But Zarkadakis continually warns us not to assume the analogy proves identity.

Early artificial intelligence

Researchers then explicitly attempt to reproduce intelligent behavior.

Early AI emphasizes symbolic reasoning.

Knowledge is represented through symbols and rules.

For example:

IF condition A + B + C → conclusion X

This works beautifully for limited domains.

But the real world is messy.

The combinatorial explosion

A machine trying to consider every possible possibility quickly faces an astronomical number of combinations.

Chess provides a simple example.

Every move creates many possible future moves.

The possibilities explode.

Humans somehow make useful decisions without exhaustively searching every possibility.

This becomes one of the great challenges for symbolic AI.

Expert systems

One solution is to encode human expertise.

An expert system might contain thousands of rules:

IF symptoms X and Y → consider diagnosis Z.

This worked surprisingly well.

But it revealed the knowledge acquisition problem.

Experts know much more than they can explicitly articulate.

Much human knowledge is:

tacit,
contextual,
embodied,
intuitive.

You cannot simply ask a master chess player to write down every process their brain uses.

AI winters

Early AI made enormous promises.

When those promises were not fulfilled, funding and enthusiasm collapsed.

These became the AI winters.

The historical lesson is important:

AI repeatedly advances through cycles of extraordinary optimism and painful correction.

That should make us cautious about both AI hype and AI pessimism.

Connectionism returns

Eventually researchers increasingly turned back toward neural networks.

Instead of:

program every rule

the new philosophy became:

let the system learn patterns from examples.

This produces:

data → learning → representation → prediction

rather than simply:

rules → calculation → answer.

Darwin at the Edge of Chaos

Zarkadakis then connects computing with Darwinian evolution and complexity.

Darwin's great insight was that extraordinarily complex organisms can emerge without someone individually designing every feature.

Through:

variation → selection → reproduction → adaptation

complexity accumulates.

This inspires:

genetic algorithms,
evolutionary computation,
artificial life,
emergent systems.

The central insight:

Complex intelligence may emerge from simple interacting processes.

Complexity and the edge of chaos

Complex systems can exist between:

rigid order

and

complete randomness.

Too much order:

→ nothing new happens.

Too much randomness:

→ no stable structure survives.

But between them:

→ patterns emerge
→ systems adapt
→ complexity grows.

This is the edge of chaos idea.

It offers one possible explanation for how surprisingly sophisticated behavior can emerge from relatively simple rules.

Moravec's paradox

AI also reveals something wonderfully strange about human intelligence.

Things that seem intellectually difficult to us—like chess or enormous calculations—can be relatively easy for computers.

But things that seem effortless—like:

walking,
recognizing a familiar face,
understanding context,
navigating a room,
picking up an object—

can be extremely difficult for machines.

Why?

Because those "easy" abilities contain millions of years of evolutionary refinement.

What is easy for a human may be computationally extraordinarily complex.

Deep learning

Modern neural networks change the picture again.

Instead of engineers manually defining every feature, systems can learn increasingly abstract representations from massive amounts of data.

Very roughly:

data

↓

patterns

↓

representations

↓

higher-level features

↓

prediction/action

This is the intellectual foundation of much contemporary AI.

The machine is no longer simply following a human-authored rulebook.

It is discovering statistical structures in experience.

And then comes the great unresolved question

At this point, we have machines that can increasingly:

perceive,
predict,
translate,
generate language,
recognize patterns,
learn,
plan,
create,
adapt.

So:

Have we built minds?

The book does not give us an easy yes.

Because the word mind hides several different things.

Five different questions hiding inside "Is AI intelligent?"
1. Can it calculate?

Obviously yes.

2. Can it learn?

Yes.

3. Can it reason?

Increasingly.

4. Can it behave intelligently?

Certainly.

5. Does it consciously experience the world?

We don't know.

And that fifth question is fundamentally different from the first four.

The Islamic perspective

This is where the whole book becomes especially interesting through an Islamic philosophical lens.

Islamic thought does not reduce the human being to a single mechanism.

Several concepts are important:

Nafs

The self—the individual moral and psychological self.

Aql

Intellect/reason—the capacity for understanding and discernment.

Qalb

The heart—not merely a biological organ, but a locus associated with perception, understanding and moral orientation.

Ruh

The spirit.

And the Qur'an gives us a striking reminder concerning the ruh:

“You have been given of knowledge only a little.”
Qur'an 17:85

That provides an important epistemological humility.

We can investigate the brain.

We can model intelligence.

We can reproduce behavior.

But we should not assume that successful modeling means we have explained the deepest reality of consciousness.

Human intelligence versus machine intelligence

An Islamic perspective can therefore distinguish between:

information

knowledge

intelligence

wisdom

moral responsibility

spiritual existence

These are not necessarily the same thing.

An AI may eventually contain more information than any human being.

It may outperform us in mathematics.

It may reason faster.

It may generate beautiful prose.

But:

Information is not wisdom.

And:

Optimization is not morality.

And:

Simulation of emotion is not necessarily experience of emotion.

The problem of purpose

This is perhaps the most important bridge between Zarkadakis and Islamic thought.

A machine can be extraordinarily good at answering:

“How do I achieve goal X?”

But that does not automatically answer:

“Is X worth pursuing?”

That is the distinction between instrumental intelligence and wisdom about ends.

Modern AI alignment often asks:

How do we align AI with human values?

But before we can answer that, we must confront a deeper philosophical question:

What are the right values?

Technology cannot simply calculate that answer for us.

Amanah: intelligence as responsibility

Islam introduces the concept of amanah—trust and responsibility.

That gives us a powerful way of thinking about AI.

The question is not merely:

Can humanity create increasingly powerful systems?

It is:

What responsibilities come with possessing that power?

This changes the technological ethic from:

“Can we build it?”

to:

“Should we build it, how should we govern it, and who bears responsibility for its consequences?”

That is a very modern question expressed through a very old moral framework.

The deepest paradox of the book

And now the three parts come together.

Human beings began by imagining artificial creatures.

Then we tried to understand ourselves.

Then we began constructing machines that reproduce fragments of what we call intelligence.

And now we are approaching a remarkable historical paradox:

We may become capable of creating increasingly intelligent systems before we fully understand the intelligence we are trying to recreate.

That is the real philosophical tension running through In Our Own Image.

The entire intellectual journey in one map

Ancient myths of artificial life

↓

Automata

↓

Mechanical explanations of life

↓

Descartes: mind/body

↓

Hobbes: mechanical thought

↓

Cybernetics: feedback + control

↓

Brain as information-processing system

↓

Computational theory of mind

↓

Turing: machine intelligence

↓

Babbage + Lovelace: programmable machines

↓

Gödel: limits of formal systems

↓

Shannon: information

↓

Von Neumann: stored-program computer

↓

Symbolic AI

↓

Expert systems

↓

AI winters

↓

Connectionism

↓

Neural networks

↓

Evolutionary computation

↓

Complexity + emergence

↓

Deep learning

↓

Modern AI

↓

The unresolved problem of consciousness

↓

What exactly is a human being?
And this is where the book becomes bigger than AI

Zarkadakis is ultimately not telling us:

“Machines will become human.”

Nor is he simply telling us:

“Machines can never become human.”

He is showing us something more intellectually interesting:

Every time humanity invents a new technology, we use that technology as a metaphor for ourselves.

The clock made us think mechanically.

The steam engine made us think dynamically.

The telephone made us think in terms of signals and communication.

The computer made us think computationally.

The Internet made us think in networks.

AI now makes us think in ter

[Florin Pitea · Rating 5 out of 5]

Very well documented, synthetic and crystal clear. Highly recommended.

[Nick · Rating 3 out of 5]

A fair book that takes a look at AI mainly through a historical and philosophical lens. The author contends that AI cannot replace human at a core layer.

[Ami Iida · Rating 3 out of 5]

Star 3.5☆☆☆+0.5
This document is information compression manual of science
Artificial intelligence is made up on the top of such as quantum mechanics, mathematics, programming, physics.

the book theme is "What is consciousness?" ans " Can A.I, have consciousness ?"
Do Androids dream of the electric sheep? LOL

the author compares artificial intelligence and science fiction and the movies, ex. 2001 A Space Odyssey ,blade runner,terminator,Frankenstein,etc...............

reading through half of this book now.
It is written many metaphysical topics
,then mind-brain problem , can artificial intelligence acquire consciousness?
is artificial intelligence going to overtake the human brain
-------singularity ---------it guesses 2045 etc.............



It is written in detail mind-brain problem.
It has also been cited opinion of renists and physicists,philosopher,etc..........
"What is consciousness?"
"Can artificial intelligence have a consciousness?"
"Singularity" is written in detail in it.

[Rahul Narayanan · Rating 4 out of 5]

In our own image - journey of human psyche till AI rather than answering an Yes or No question ,rightly so.to know how we are coping to end up or even start life continuum with AI conundrum ,george zarkadakis takes us to examine theory of mind ,mind body question , evolution of language and its components, with pinch of history on thoughts behind programming evolution from 17th century till 21st.a fairly good start for anyone interested in AI even though last few chapters struggles with the focus and left me with scattered thoughts on a diluted ending.

[Ryan · Rating 4 out of 5]

3.5, a great run through of the history, science, and philosophy of AI. Gets a bit extended into speculation beyond the titled scope and there is certainly no subtly in the authors bias, yet definitely recommended for anyone i interested in the area.

[Bryce Van Vleet · Rating 3 out of 5]

A fantastic, if a bit dry, book that covers AI from every lens imaginable - anthropology, literature, neuroscience, and history. This book is a fantastic resource for those who find themselves captivated by a future increasingly becoming mechanical.

[Jared White · Rating 5 out of 5]

Fantastic philosophical, historical, and scientific treatise of AI! I highly recommend if you want to gain a firm grasp on the underlining philosophical and technical underpinning of AI development and barriers to advancement.

[Tom Beck · Rating 5 out of 5]

I great overview and summary of everything I’ve read on AI

[Chris Esposo · Rating 5 out of 5]

Although the book is advertised as a tour of the concept of AI with the headline areas being fiction and its media, and other public forum dimensions, much of the book is actually a fairly deep look at AI from the view of formal computing, biology, neuroscience, mathematics, and philosophy. Thus, the book is much more of a serious attempt at synthesis of all the above to answer: What is AI, how could it be achieved, and what are the likely impacts it will have on society going forward. Though unlike other more pop-sci books, this one focuses much more on the "what?"

In particular, the book builds up in logical theory through a combination of a fairly detailed exposition of results and historical stories to reach the point of formal incompleteness, and what that means with respect to assigning true/false values for self-referential statements in deduction systems. The author seems to imply that any true AI must be able to master self-references, not falling victim to the "Halting Problem" that is logically equivalent to incompleteness.

The author makes a feasibility argument that sapient agents seem to be able to grok self-reference (at least the living organisms we can observe on Earth), and thus, it is at least a necessary condition for AI. This is an "argument by qualia", and the author also later introduces the more practical definition of intelligence utilized by CS known as the bird/airplane analogue argument. This analogue is a utilitarian approach, showing that there are multiple avenues to achieve one 'property', where one need not simulate all sub-systems of an agent that has this property. By example to gain the property of flight (or possess super-flight relative to a bird), we need not recreate the exact mechanics or morphology of a bird wing, but just a simple abstraction of that wing, to get to a device that is mechanically superior to the original with respect to speed and durability. Likewise, we need not recreate anything close to the exact neural network of an animal mind to build super-intelligence.

The issue of assignment of a value for self-referential introduced by the author is subsumed in the more general notion of the "Entscheidungsproblem" of Hilbert in formal computing, and the author suggests that a stochastic value assignment could be a way out of this mess of resolving the problem that will get us closer to the "what?" of AI. In AI, a system that can assign values to these self-referential and still maintain internal-consistency (each marginal assignment do not generate contradictions), is referred to as "Turing Complete". As an aside, the search for finding the right assignment procedure for these class of statements is different from the path mathematicians took historically, which was to preclude any such statements from entering a formal calculus by refining admissibility on the set of well-formed statements for such a calculus.

Ultimately the author's goal by all this work is to connect the notion of self-reference to notions of cybernetics. This is where it gets hazy, and will probably require a review or guided second reading, as the exact nature of the connection eluded my memory after my first go-through, except a vague notion that cybernetic systems include ideas of system dynamics where the idea of "feedback loops" is central to understanding nonlinear behaviour, and these could be viewed as self-referential.

As one should be able to tell the book is fairly dense with ideas, and those mentioned here are only from one thread. The author covers much material on neuroscience, as well as a deep dive on the mind-body.problem of Descartes, in some ways the foundation for the central theme of the text. Besides the academics, about 20% of the text is devoted to AI in fiction and stories, though mostly from human anthropology with a choice selection from some well-known fiction like Frankenstein and other sci-fi authors.

Though not what I expected, what one gets is a very substantive discussion on AI, serious and deep enough to be included in a reading list for a special topics course on the subject, at either a graduate or undergraduate level, as it definitely adds color to the standard technical treatment of the subject. Specifically for those who are or planning to specialize in this field or something adjacent, I found it especially insightful for perspective on the history of AI and parts of ML for those who joined the field post the first AI-winter. Really great book. Useful. Highly recommended.

[YHC · Rating 4 out of 5]

This book is in 3 parts.
First about the homo sapiens history, we mostly know the development and evolution about human history. Second part about robot development including the computer. Third part is about the consciousness on human and robot. This part is more fascinating. The definition of AI is artificial intelligence. But how do we define intelligence on Humans? Our behaviors, our morality, justice, doing the right things, are we so sure about human are intelligent?

We are talking about artificial intelligence on robot with human kind consciousness, which doesn't exist so far and not able to find similar kind among any other animals except humans. We surely capable to use computer to replace many human fastidious or boring jobs, but the consciousness is not given. The author give us a living mechanism such us the whole earth ecosystem itself has no consciousness.

Therefore he didn't jump to conclusion about what if the robots could actually got chip in our human intelligence (with the touch of consciousness) since we are not there yet.

With recent on line course i learned on "element of AI", it explained in a more neutral way about if AI is going to be dangerous and wipe out human being, again similar that we wish to use robots to facilitate our life, not to replace us.

[Nujood AlMulla · Rating 5 out of 5]

I can’t begin to explain how much I enjoyed reading this book. The first two parts specifically were groundbreaking for me personally as I am studying a specific area in AI but never came across an interpretation of its history in the way Zarkadakis put it. The links to dualism and philosophy were described very clearly and systematically. The novels and fictional characters he included and how they were linked to our current aspirations in AI was inspiring. Not to mention his comprehensive explanation of the moments the pursuit of an artificial mind changed from a holistic recreation of the human mind to a more grounded mechanical role. I highly recommend this book to anyone in the field.

[Neil H · Rating 5 out of 5]

Not a short book by any means but chock-full of philosophical and science history that provides a good narrative to what we consider as intelligence and artificial intelligence. Questioning the largely acknowledged fact of the inevitable march towards progress. We see that all our achievements has not been linear. In fact our theories, execution and possibilities of variant philosophical and scientific to have come to this is by a long shot. I'll come back to this book again and again.

[fff · Rating 4 out of 5]

Many good points raised here, but most of them are common-sensical ideas which most readers already understand. This book raises more questions than answers. The arguments presented here rest more on philosophical logic and inquiry rather than computer science or examples, which was disappointing. This wasn't what I hoped for when I picked up this book.

[Michael Brenner · Rating 2 out of 5]

This seems to have been well researched and it was an interesting read. That said, the overt Marxism, materialism, and trans-humanist agenda being pushed was assumed, not defended. From a Christian standpoint, the book’s message was garbage.

[Aca · Rating 3 out of 5]

Long. Overview of sociology, neuroscience, AI

AI is created by logic limited to human perception