Hannu Toivonen releases a popular science book on artificial intelligence

With artificial intelligence remaining the subject of heated debate, the topic should be opened up to the public in Finland too. Professor Hannu Toivonen has published a book that does just that, while also exploring misunderstandings and misconceptions about AI.

Professor of Computer Science Hannu Toivonen is an expert on AI and computational creativity. Having previously commented on the topic on television and radio, Toivonen has now published a popular science book entitled Mitä tekoäly on – 100 kysymystä ja vastausta (‘What is artificial intelligence – 100 questions and answers’, Teos 2023).

“People seem to be yearning for information on AI right now. I wrote the book to explain the interesting and important issues surrounding the topic to the general public and tackle some prejudices and misunderstandings,” he says.

Although the issues are of global significance, Toivonen found it important to write a book in Finnish for Finns.

“If it had been a scholarly book, I would have written it in English. But I think it’s best to write a popular science book in your own language to be able to express yourself in a more nuanced way. I wrote this book to promote Finnish AI literacy, and I believe it’s easier to do that with a book written originally in Finnish rather than a translation,” he notes.

The book is written in a popular and accessible style, using few technical terms. The numbered structure makes it easy for the reader to browse the topics and issues discussed.

“AI is a broad field, and the book only scratches the surface. To create a reader-friendly book, I didn’t even try to make it comprehensive, as that would have required a whole series of books – or more,” Toivonen says.

“Instead, I handpicked 100 topics I wanted to write about for Finns. My aim is to provide a diverse view of the field, which is why I’ve chosen as wide a range of issues as possible.”

AI models available for public testing

Toivonen started writing the book in the summer of 2022. Since then the AI debate has evolved considerably. ChatGPT and other language models and applications such as image generation are now widely available to consumers.

“Almost anyone can use the kind of AI software that was unavailable even to researchers just a short while ago,” he points out.

Toivonen himself has a long background in research on AI creativity. The use of AI for creative work is a source of controversy that has contributed to the Hollywood writers’ strike, for instance.

“The use of AI in the creative sector is new. But automation has been revolutionising work for centuries, so this latest labour dispute is not the first of its kind,” he says.

It has been estimated that the development of AI puts up to one-quarter of all jobs at risk.

“But people can never be fully replaced. We are needed to develop and use AI software because we can understand and appreciate the wider context.”

Misunderstandings are common

Toivonen says that some issues in the AI debate have been sensationalised. This may have given a false impression of the autonomy of AI.

“AI doesn’t, for instance, ‘crack our passwords’, as has been claimed in some newspaper headlines. It’s we humans who use computer software to crack passwords. The impression often given of AI software is that of an active agent, which just isn’t true,” he notes.

As AI appears to do the same things as humans do, we may mistakenly believe that AI is able to think like us.

“Language, for example, is the way people communicate their thoughts to each other. So when a machine uses language, we believe it too has thoughts to communicate, even though it functions on the basis of statistical language models rather than understanding the meanings of words,” says Toivonen.

“Likewise, the generation of images on a given topic may give the impression that the program understands the content of the images it produces.”

Toivonen mentions the common misconception that computer programs are unbiased and neutral.

“A computer program always reflects the values of its creator and, in the case of machine learning, the values and attitudes of the material used to teach it. We have lots of examples of discrimination by AI software.”

Principles endure

Because the book looks at AI from different perspectives, Toivonen has discussed its content with experts from other fields.

“It’s been interesting to explore issues such as awareness, language, law and power. I’d like to thank my colleagues for their support. I hope I haven’t oversimplified things,” he says.

Although AI is advancing in leaps and bounds, Toivonen believes his book will remain relevant for years to come.

“It’s not about individual programs or technical solutions, but about principles and concepts, and principles change slowly. New issues may of course arise quickly. Then I’ll just have to write a new book on the next 100 questions and answers,” he states.


The book launch will take place on 13 October at the Finnish Science Centre Heureka.