Karen Hao: Empire of AI (2025)

"Empire of AI", by Karen Hao, tell the history of OpenAI, from its creation as a non-profit with the ambition of developing AI for the benefit of humanity, to the profit-making company behind ChatGPT; one of the most powerful actors in AI today. 

Much of the book follows Sam Altman and the people around OpenAI: It describes their disagreements, changing business model and eventually the crisis in 2023, when Altman was fired by the board only to return a few days later. This history is interesting because OpenAI is not just another company: it is influencing how one of the most important technologies of our time is developing. 

I felt that the book at times became too personal (some of the conflicts sound very similar to what can happen in any office…), and some of the biographical material unnecessary for the book’s point (in particular on Mr. Altman’s sister). 

What I found much more interesting though, were the disagreements about what AI actually means and where it may take us. In fact, OpenAI itself was born out of these questions: If more powerful AI can transform society, who should control it? Should development be accelerated because of the enormous benefits it could bring, or slowed because we do not understand the risks? The disagreement between the AI “boomers” and “doomers” runs through much of the history of OpenAI and helps explain many of its internal conflicts. Thus, the best part of the book is when it looks beyond OpenAI itself; for instance when it discusses the enormous data centers needed to develop AI (including their consumption of energy and water, and the human labour behind AI, including workers in developing countries hired to label data and moderate disturbing content). The apparently immaterial world of AI actually depends on very physical infrastructure, resources and human labour, something that is too often lost in the wider discussion on the technology. 

Ms. Hao’s main argument is that there is an emerging “Empire of AI”; a small number of powerful tech companies and individuals concentrating technology, capital and decision-making while exploiting resources, data and labour from across the world, just like any previous capitalist exploitation throughout history. 

I was less convinced about this argument, as I think that it on the one hand ignores the complexity and speed of technological change, as well as simplifying the social dynamics of such change. I got a bit annoyed when she appeared to criticize people at OpenAI for not providing satisfactory answers about the consequences of AI, implying that they were avoiding difficult questions. I rather think there may be  another explanation: they simply do not know the answers! 

And that was indeed the most interesting conclusion I took from the book: The people developing AI are obviously intelligent, ambitious and visionary, but they are all still human beings! As Ms. Hao amply proves, they disagree, make mistakes, fight over office politics and interpret evidence differently, and in fact these people themselves do not really know how quickly it will develop, what its impact will be, what resources it will require, what it will do to employment and inequality, or what role it will eventually have in our societies. 

I therefore think the book could have been much stronger if it had used the history of OpenAI more explicitly as a way to explore these questions. 

Maybe my slightly uncomfortable conclusion from the book is that there appears to be nobody anywhere who fully understand what is happening with AI: Not OpenAI, not academics, not governments, not business... They have different ideas about what might happen, but nobody really knows where AI is going or how quickly we will get there... 

Of course this does not mean that we should panic about AI, nor that we should simply leave its development to tech companies and billionaires. It rather means that we should talk much more seriously about what we want from this technology, what risks we are willing to accept, how its benefits and costs should be distributed, and what role we want it to have in our societies. 

That is also why I would still recommend the book as one more in a canon of AI books; read together with other AI books it shows the very human organisation behind a technology that can otherwise seem almost abstract, and thus the questions we are all struggling with.

Nigel Toon: How AI Thinks (2024)

This is a great book, very up to date (2024) that basically aims to explain how artificial intelligence has developed, how it actually works, and what its impact on society may be. It thus covers all the discussions and is written for a general audience, taking on the task of demystifying AI at this moment when the technology is coming to the fore in all aspects of our lives. The book’s message is that AI is neither magic nor science fiction; it is rather a powerful tool built from data, algorithms, and computing power—capable of transforming society, but still fundamentally different from human intelligence. 

The book begins by tracing the origins of AI; a story we have heard before, but that focuses on the shift from programming of logical machines, to machine learning, where computers learn patterns from large amounts of data instead of relying on explicit rules. This change, combined with stronger computing power (he mentions that computing power to train AI models increased 300 000-fold between 2012 and 2019), as well as the growth of massive datasets, have led to today’s AI breakthroughs. An important points that Mr. Toon tries to get through throughout the book is that modern AI systems do not think like humans and their so-called “intelligence” is rather limited from its lack of experience and intuition. He quotes Alan Turing in a short chapter that tries to answer “what is intelligence”: “If a machine is expected to be infallible, it cannot also be intelligent”. 

Instead, the AI systems learn statistical patterns. Neural networks analyze enormous quantities o if examples—images, text, or speech—and gradually adjust millions of parameters until they can make probabilistic predictions or generate new outputs. It is a system approach that can complement human intelligence, but not replace it. 

Following the development of AI and how it works, the book then looks at the societal impact of AI, arguing that AI will transform many industries—medicine, science, finance, and creative work. However, Toon warns that people often misunderstand its capabilities. AI excels at pattern recognition and prediction, but it does not possess consciousness, understanding, or human common sense. Because of this, AI can still make strange mistakes or produce incorrect information even when it appears “intelligent”. The author still finds that the opportunities of AI are enormous in terms of accelerating scientific discovery, improving healthcare and climate research and automating complex tasks; these are areas where we are already starting to see the impact. At the same time, he also stresses the need for governance and control, ensuring that AI systems are used responsibly and that society understands how they operate. AI is a tool, and in the end it is about how human uses them. 

A great book, perhaps one of the best to read currently on the subject for general audiences.

Kai-Fu Lee & Chen Qiufan: AI 2041

“AI 2041” by Kai-Fu Lee and Chen Qiufan is a fascinating mix of storytelling and analysis, imagining how artificial intelligence could shape our lives over the next two decades. The book explores AI themes though ten short stories set in different parts of the world, and then follows each story with an essay explaining the technologies behind the stories. 

The central message is that AI will not arrive in one sudden leap, but through gradual integration into almost everything we do. 

Having been written in 2021, the book looks 20 years ahead, imagining that by 2041, AI will be integrated into much of our lives — in healthcare, education, finance, transport, entertainment and public services — shaping decisions, and entailing opportunities and risks that will deeply transform our societies. 

A recurring theme is the geopolitical and economic dimension of AI. Countries with access to large amounts of data, computing power and talent will lead the race, with the United States and China clearly positioned at the forefront. This growing concentration of technological power could reshape global influence and economic competition. For other countries and regions, including Europe and the Global South, the challenge will be how to remain relevant. 

On the societal side, the book treats different subjects, from love to how AI will reshape work. Many routine or repetitive tasks are likely to be automated, from logistics and manufacturing to administrative functions. The short story “The Job Savior” is about a man whose job has been automated works as a “re-employment companion,” helping other workers adapt to AI-driven changes and reflects on automation, large-scale job displacement, and the psychological dimension of this ongoing economic transition. 

However, the book does not adopt a purely pessimistic view. New forms of employment will emerge around AI oversight, human-centered services, creative industries and system design. The real risk lies in the transition: without re-skilling, education and social policies to cushion the shift, inequality could widen both within and between societies. The short story “The Golden Elephant” is optimistic in being about a rural Indian girl uses AI-powered education tools and gig platforms to lift herself out of poverty. This is in contrast to the more pessimistic story "The Holy Driver" about the dominance of autonomous vehicles, making human drivers obsolete, except for those who resist. 

Another important thread is human–AI collaboration: The most plausible future is not one in which machines simply replace humans, but one in which they augment human capabilities. Doctors supported by diagnostic algorithms, teachers assisted by adaptive learning tools, or AI systems working with lawyers to strengthen access to justice, are presented as possible scenarios in different stories. In this sense, the book shows that the societies that can benefit most from AI will be those that manage to combine technological efficiency with human judgement, empathy and creativity. 

The book does not shy away from many of the ethical concerns that AI is already raising today. Issues such as surveillance, algorithmic bias, deepfakes and concentration of technological power run throughout the stories. The story “Quantum Genocide" is about the development of quantum computing that breaks all encryption standards, triggering global cybersecurity chaos; the story is a stark warning of the risks of some breakthroughs and the fragility of digital infrastructure. The story “Deepfake Justice” presents a world where AI-generated deepfakes have become so sophisticated that truth itself is contested and legal systems struggle to determine authenticity as visual evidence can no longer be trusted. 

This is a great and accessible book on AI. I don’t only like the approach of the stories, but also its balance: AI is not inherently good or bad - it carries risks as well as potential, and its impact will depend on governance, regulation and the values embedded in its design and use.

 

Erik Brynjolfsson and Andrew McAfee: The Second Machine Age

The Second Machine Age by Erik Brynjolfsson and Andrew McAfee is a book from 2014, and perhaps one of the first books that tried to dwell into the impact that digital technology will have on the economies, and particularly on the nature of work and employment. With the advent of AI technologies this discussion has become even more important, and while they do not specifically go into AI as such, the book remains highly relevant today. 

The book's parting argument is that hat humanity has entered a new era in which digital technologies (eg. computing power, artificial intelligence, robotics, and networks) are transforming economies and societies as profoundly, -perhaps even more -, as the Industrial Revolution once did: this is what they call the “second machine age”. 

Brynjolfsson and McAfee argue that technological progress is accelerating exponentially, driven by improvements in processing power, data availability, and connectivity. This progress enables automation of routine and even complex tasks, boosting productivity and creating enormous wealth. However, there are great risks that the benefits are unevenly distributed: while innovation can generate growth and new opportunities, it also risks widening inequality, displacing certain jobs, and concentrating rewards among highly skilled workers and technology owners. 

Brynjolfsson and McAfee make an argument that in my view is highly relevant today: technology itself is not the main problem; rather, the challenge lies in how societies adapt. This we are indeed seeing, as the technology of AI appears to move at a speed where we are unable to adapt, and this causes both fear and concern in many of us, that the skills we have used lifetimes in learning, become useless in the blink of an eye. The authors of this book call for updated institutions, education systems, and policies that foster innovation while supporting workers through transitions. Encouraging entrepreneurship, investing in digital skills, and rethinking regulation and social safety nets are central to ensuring that technological advances improve broad well-being. 

Ultimately, the book is inherently optimistic in that it argues that the second machine age offers unprecedented potential for prosperity—if societies manage the transition wisely, and this appears even more relevant today than when the book was written, and this is a book that stands as one of the best early books on this subject.

 

Ethan Mollick: Co-intelligence -Living and Working with AI (2024)

The book is written by Ethan Mollick, a leading expert on AI from Wharton University. The book is a relatively short read, written as a practical and optimistic guide to understanding and working with generative AI. The author focuses on how AI can work with us humans: instead of seeing AI as a threat or a distant future risk, Mollick argues that AI should be seen as a collaborative partner (“co-intelligence”) that can enhance human creativity, productivity, and learning. 
The book explains how large language models work, why they matter, and introduces key principles for how people can work with AI:
Always invite AI to the table: Even when you think it won’t help, try it. AI often adds unexpected value — ideas, structure, alternatives, or speed, and it can help you in terms of developing ideas.

• Be the human in the loop: No matter how you use it, you are the boss. Humans must set goals, context, and judgment. AI supports decisions; it should never silently make them.
• Treat AI like a person (but tell it what kind of person it is): He says that while not anthropomorphizing AI, the richest interaction is when you start treating it as a human. In this interaction you will develop how to to prompt. Asking good questions, giving examples, and refining instructions is now a core skill.
• Assume this is the worst AI you will ever use: Never fully trust AI, as it can sound confident but still be wrong. Treat it as a brilliant but unreliable collaborator. Verify facts and logic, but keep experimenting and challenging it.

All in all, Mollick is an optimist about how we are able to use AI in our everyday life, and I have personally started using some of his principles as I experiment with an AI assistant in many everyday tasks, impressed by the technology, but always keeping in my head that it can and should not replace human judgment.
The book is strong as a guideline for everyday use, but not on the societal impact of AI. Although implicitly Mollick is an optimist, he fails to address issues such as unequal access, the impact on jobs (he is clearly of the view that it will not displace jobs for humans, but create new necessities), market concentration, and how AI may come to affect identities and cultures as well as the possible erosion of human enterprise.
It is a good quick read, specially if you want to start experimenting with AI in your everyday life, but it should not stand alone, in particularly on debate on the social impact of AI.

 

Chief Hatuey

 In the 16th century, the Taino Chief Hatuey valiantly resisted the Spanish conquistadors when they arrived to his native island that today is known as Hispaniola. When the Spaniards almost caught him, he fled to Cuba, where he prepared local natives to resist the coming Spanish invasion.
And when the invasion came they bravely attacked the Spaniards, who had to use all their cunning, technology and vicious hunting dogs to slowly but steadily destroy all native resistance. Chief Hatuey was caught and was immediately sentenced to die a slow painful death by burning; this was the worst punishment that the Spaniards would give to criminals and heretics. 
So it was therefore that a Priest approached Chief Huatey:
“If you accept the true lord Jesus Christ as your Saviour we will give you a quick death”
As the Taino Chief, bound and wounded, failed to reply the Priest continued trying to convince him:
“If you accept the Lord Jesus Christ you will go to Heaven!” 
Chief Huatey looked at the Priest with repulsion as the words were translated into his native Taino language.
“Do the Spaniards go to Heaven as well?”, he asked the Priest.
“Yes!”
“Burn me.”

Max Texmark: Life 3.0 (2017)

Max Tegmark is a Physics professor at MIT, who wrote this book already in 2017, when we were still to see some of the extraordinary advances in AI that we have seen in the last few years, specially with the advent of ChatGPT. I found this book an excellent expansive and curious exploration on what AI could mean for human life, starting from the physics of the early universe to speculative visions of superintelligent minds, and how it might develop with. While highly speculative, the book is not naive, rather opening up a lot of questions both on evolution, the nature of "intelligence" as well as visions about the future. 

Life 3.0 opens with a simple question: what happens when intelligence decouples from biology? The author then takes an evolutionary approach to frame how organisms evolve, learn, and ultimately redesign themselves. He then moves through the technical foundations of AI, exploring how current systems work and what “general intelligence” might actually mean. 

After this introduction, the middle of the book widens into almost- science-fiction-like future scenarios, some utopian, some bleak, and some delightfully weird, always prompting the reader to ask not just what could happen but what we would choose. The final part of the book focuses more on governance and values — how to align advanced AI with human goals, how institutions might adapt, and how to avoid the classic traps of short-termism and fear. 

Life 3.0 is a philosophical and strategic reflection in thinking about the long-term future of intelligence. Many of its big-picture arguments remain extremely relevant and wise. At the same time there are some weaknesses, based on how we have seen AI evolve over the years: Tegmark focuses a lot on an AGI evolving rather “suddenly", but today the trajectory looks more like a continuum of capabilities scaling iteratively across domains, rather than one sharp inflection point. This is also something that influences the development of our values and institutions in the face of the technology: it is evolving in parallel, as we adapt to the technology; a technology that is much more embedded in society today that I think even Tegmark had imagined when writing the book. 

But a great book, reflective and not too technical. What I liked most about the book is its high level of reflection, that connect evolution, intelligence, values and governance, but also uses imagination and creativity. Despite some bleak scenarios, the author is generally optimist but cautious, urging us not to sleepwalk into our future, but to arrive at it with curiosity, responsibility, and a bit of courage.