In his very first sentence of the preface, Toby Walsh warns us “This book is out of date”. He’s right- it was published in 2023 and it does already feel out of date. In 2018 Walsh wrote a book called 2062, which was the year that he predicted machines would be as intelligent as humans. I wonder if he still feels that way. In reading Faking It in 2026, three years after it was written, it feels as if things are moving much more quickly than he has described them. But, also as he says in the preface
I am confident, however, that the issues raised in this book will not be out of date. Indeed, they will be even more pressing. And this book will be even more useful, as a guide and a warning.
Toby Walsh is Scientia Professor of Artificial Intelligence at the University of New South Wales in Sydney and CSIRO’s Data61. He is a scientist who embraces the potential of AI but has been warning the government of the dangers of unregulated AI for some time as, indeed, he also does in this book. I feel that he takes a balanced approach, neither warning of impending doom but neither is he cavalier about the potential dangers that it could bring.
As suggested by the title, he is adamant about the ‘fake’ nature of AI.
In understanding artificial intelligence, it is also easy to be misled by our own human intelligence. We have a very personal experience of our own human intelligence, so it is hard not to fall into the trap of thinking that artificial intelligence might be somewhat similar.
Much of this turns on how we describe intelligence.
What, then, is intelligence? Loosely speaking, intelligence is the ability to extract information, learn from experience, adapt to the environment, understand and reason about the world. So what, you should ask, is artificial intelligence? Most AI isn’t embodied and situated in the world like our human intelligence, adapting and learning from the environment. How, then, can the intelligence of machines be identified and measured when it is so fundamentally different to human intelligence?
AI researchers do not completely agree on how to define artificial intelligence. But most of us will agree that it is about getting computers to do tasks that, when humans do them, we say require intelligence. These include perceiving the world, reasoning about the world and learning from the world.
His main argument throughout the book is that the artificial intelligence we are building only simulates human intelligence, – in other words ‘faking it’.
This fakery existed right from the start. The Mechanical Turk, a chess playing automaton which debuted in 1770 and toured Europe and the Americas until its destruction in a museum fire in 1854, actually had a person concealed within its metal chest. Mephisto, another chess-playing machine from 1876, was directed from another room using an electro-mechanical connection. Even Alan Turing faked his first chess-playing program, simulating the calculations of the program with pen and paper. And when Elon Musk announced the yet-to-be built Telsa Bot in 2021, it featured a person dressed up in a white full-length bodysuit pretending to be a Tesla Bot and fooling nobody.
He identifies many situations where artificial intelligence is much less impressive than it appears, and where it is designed to deceive us. He starts by looking at the hype and false claims about artificial intelligence; then looks to how AI has tried to fake human intelligence; to create fake real people, and to emulate human creativity. He notes that the technology companies that are developing AI are often indulging in fakery themselves, often indulging in ‘ethics-washing’ that makes them sound more responsible than they are.
In describing ‘the artificial in AI’, he reminds us that psychologists distinguish between intellectual, emotional and social intelligence. AI has the first one covered, but the other two intelligences are less easily replicated. AI machines, as strictly electrical devices, do not have the biochemistry that drives our emotions. Of course, they could be programmed for a level of anger, disgust or fear but our emotional intelligence is informed by our lived experiences: we understand other people’s emotions because they are like our own. Social intelligence, he suggests, is perhaps our greatest superpower. We can empathize with other people because we share the same biology. This, too, can be programmed, but that assumes that people act rationally and in self-interested ways, ignoring the complex, often irrational and sometimes altruistic behaviours of real people.
He urges those working in the field to stop anthropomorphizing the technologies they are developing:
We talk about a chatbot ‘understanding’ a sentence, a ‘self-driving’ car, the computer-vision algorithm ‘recognizing’ the pedestrian, and the possibility of robot ‘rights’. In reality, chatbots don’t understand language. There is no self– no person, no sentient, self-aware intelligence, driving the car, even if the car is driving itself. Algorithms don’t actually recognize objects. And robots need rights about as much as your toaster does.
His final conclusion is that the human in human intelligence is perhaps the most important distinction between us and the ever more capable and artificial machines we are building. He reminds us of the COVID times, when we so much missed physically meeting people, and walking in the real world. It was the real world we craved.
This is a very convivially-written book that is packed with information, and yet it rarely feels overwhelming (with the exception of his examples from mathematics which challenged my non-mathematical human brain). His footnotes are quirky and humourous and he has a wealth of examples and anecdotes drawn from his long history in the field. He is enthusiastic about the potentials of Artificial Intelligence, but adamant that it is not the same as human intelligence and in this I think that he finds the right balance.
My rating: 8/10
Read because: I’m thinking a lot about AI and being human. And it both excites and frightens me (probably more of the latter) because I want to hold on to the human.
Sourced from: purchased Kobo book (hence the lack of page numbers to the quote- a very annoying feature of Kobo)















