"I am glad that artificial intelligence is still not able to take creative work from us": An Interview with Andriy Tuzhykov

Andriy Tuzhykov is a Ukrainian writer of prose and poetry, columnist and cultural manager. He is particularly interested in artificial intelligence and quantum computing. He is the author of the book for teenagers called 'Another Brick in the Wall', and two very different books which were published this year: a collection of poems 'Тривожні маки' (Eng. 'Restless Poppies') and a non-fiction book for teenagers 'A Short History of Technology, or How to Understand Your Gadget'. We talked with Andriy about creativity, the future of literature, and discussed whether artificial intelligence could replace the writer.


You have experience in writing both poetry and prose. Was it difficult for you to master a genre of non-fiction?

It is easier to work in the non-fiction genre for me. Because when you work on your prose piece you have to create a new world: a history, characters, scenery, and you have to always watch out for the logical holes or affectation. It is way easier with non-fiction, for everything you describe exists already, and the validity of all that is created is provided by the nature itself. Sometimes it seemed to be the least creative genre of literature for me, because when I was working on the ‘The Short History of Technologies ...', I spent most of the time thinking over what I should write about, what I should avoid and how I should explain the neural network to people who do not know anything about linear algebra. The last thing I was thinking about was metaphors and images.


You received a Bachelor's Degree in Software Engineering and a Master's degree with honors in Theoretical and Computational Physics at Chernivtsi National University. Did your degrees in those fields of study influence your preferences as a writer?

They positively influenced my reading preferences. I treat classical literature with great respect, I am very much interested in contemporary Ukrainian literature, but most of all I resonate with science fiction: William Gibson, Philip Dick, Neil Stevenson, and my latest discovery is Peter Watts, who writes in the genre of hard science fiction. The highlight of this genre is that while the scientific facts described there are fictitious, they sound so convincing that you almost believe they are true. In the novel ‘Blindsight’ the emergence of vampires is explained through the mutations of the X-chromosome, and the antimatter engine is described through hypothetically possible technology. There is also a bunch of interesting facts from linguistics, AI, biology, biochemistry, neuroscience, (astro)physics and even psychology. The author writes that psychologists of the future will be able to change neural connections in the brain thus depriving the client of traumatic memories.

As for my texts, my education did not just affect the topics I work on, it is actually the reason why I write. The fact that the smartphone I use mostly for scrolling the newsfeed on Facebook is 10,000 times more powerful than the computer that controlled the first lunar landing mission inspires me greatly. I want to tell others that in fact there is no dwarf or 'fixie' in their smartphones who shows pictures. These are all bits, bytes, and programming — everybody can understand it. There is no magic, only common sense. So I began to work on a non-fiction book about technologies, and managed to finish it.

Which genres, in your opinion, does Ukrainian literature lack, and is there a demand for non-fiction literature among Ukrainian readers?

I am going to give you a very personal reply, but I feel the lack of modern science fiction. Science fiction literature is still poorly presented in the Ukrainian book market today, although this segment is gradually getting more popular. There are almost no Ukrainian authors writing sci-fi, so those few who do exist have no competition, so it affects the quality of their works and makes them worse in comparison to their foreign counterparts. However last year I could have complained about the lack of literary reportages in the bookstores of our country, and the dynamics of growth and interest in this genre is now snowballing. As far as demand is concerned, publishing is one of the few spheres of the economy where the demand can be created artificially — a publisher can shape the reader's taste and then play and satisfy his/her request. It is happening even now in the non-fiction market of Ukraine. We can see how translated literature turned from a semi-marginal segment to a widely-popular one. So now, a few years later, we have the whole series and publishing projects of documentary literature: biographies, motivational literature, popular science literature, and travel stories on the shelves of our bookstores.

DSCF6399.JPG

What feedback have you received so far for your book ‘A Short History of Technology…’ so far? 

'The Village' magazine, which included my book in their list of the top 10 books of September, says that I have made a right emphasis on the fact that technological progress is not static and what we now see as magic can very soon be put into practice. It also mentions that despite the difficult subject, the book is written in a simple and understandable language. I like this feedback because I would like my readers to perceive my book in the same way.

At my book launch in Lviv I received a question I am still looking for an answer to. The question was why bitcoins were better than the modern banking system. These questions are usually put by the blockchain systems developers, economists, and it is fine, but when ten-year-old girls ask about it, it is adorable. Ukrainian children are forward-minded and hungry for knowledge. So their participation and interest in book launches are also a kind of feedback that testifies to a request for new knowledge written in a plain language. Another question from teenagers that I liked was ‘who is smarter: robot Sophia or Apple voice assistant Siri?' It is very complicated and simple at the same time: what are the metrics of intelligence? There is still little known about intellect, so these questions alone are the reason and the driving force for the development of the IT industry. Children love such binary clashes, like ‘who is stronger: Batman or Superman, Magneto or Professor X?’ In the case of Siri and Sophia, even engineers of artificial intelligence rack their brains on this question. 

Another popular feedback is gratitude for mentioning the role of women in the history of computer technology. I do not like the stereotype that only boys can be programmers and engineers and that STEM is not for girls. If we recall the history of computer development, the first programmer was Ada Lovelace, Lord Byron’s daughter. She wrote the world's first software code for the Babbage analytical engine (even though it existed only in the form of drawings). Another example is a military woman and engineer Grace Hopper, one of the creators of a compiler, a program that turns the source code into a code the processor can understand. If it had not been for this invention, we would still write programs using machine code.


Which kind of literature is more important to read for a contemporary person: a popular scientific one or fiction literature? 

You should not put non-fiction in opposition to fiction literature. These are different ways of conveying the meaning. In simpler words: non-fiction is what has happened, and fiction is what could have happened. Fiction transmits sensory experience while non-fiction addresses the rational. Sometimes these genres intertwine, as in the literary reportage, but they are entirely different tools which are equally necessary. It is like asking what is more important: eating healthy food or doing morning exercises. Both are important.


Taking into account how rapidly technologies develop nowadays, what do you think literature of the future is going to be like? How will genres, shapes, styles and themes change in the future?

It is hard to predict, but we can try. First of all, an impetuous digitalization of various spheres of human life gave birth to cyber-punk, a literary genre that describes a hopeless future in which machines dominate people. Do you remember Gibson's 'Neuromancer' or 'The Matrix'? The emergence of tools for creating audio, video and other multimedia as well as the Internet has led to the creation of new forms, such as video poetry, interactive poetry, or graphic novel with AR-technologies. These technologies have existed for a long time now. But let's see what literary movements appeared in recent years and had not had enough time to influence literature yet. Nowadays we can see how rapidly artificial intelligence, nanotechnologies and bioengineering are progressing. Let's imagine a writer of the end of this century. He sits in a house printed on a 3D printer, eats capsules containing all the necessary substances for his body. Real food is an artifact for the rich because the planet cannot afford to feed a growing population anymore. He is not busy, because a humanoid robot does everything instead of him. What does he write about? I have no clue. Perhaps he could be inspired by nanorobots, whose painstaking and invisible work clears the world’s oceans. Perhaps he describes the grace of a robots’ movements and the taste of capsules or writes a historical novel about the illiterate humankind of the 21st century that ruined the ecosystem of the planet. Maybe he does not write it from the Earth but an orbital station or a Martian colony. Maybe he is not writing but only thinking about the text, while the neural interface is turning it into a text file. I can only fantasize on this subject, but apparently, I cannot make any forecasts.


The neural network, trained on seven books about Harry Potter, managed to write a new chapter of this story. This text was, of course, imperfect, and the plot seemed absurd. But is it possible to still improve the neural network so that it could write complete texts? If so, can artificial intelligence replace the writer in the future? 

I cite a fragment from this text in my book, by the way. But I have to warn you: do not confuse consciousness and intelligence. These are utterly different things. The neural network is just a set of thousands of values selected by the learning algorithm. The neural network sees a picture with a cat and says 'it's a cat with an accuracy of 99.9%', but it does not even realize what it is doing. How did it happen? It was shown thousands of other pictures of the cats before, so now it can recognize other cats. But once you show it a picture of a dog, the neural network will not recognize it. It will only say that it is a cat with a probability of 20%. So you need to show it another thousand images of dogs so that it could recognize them as well. But, in fact, there are no cats and dogs as such for it, there are only numbers and markers ('a cat', 'a dog'). It will probably be to the point here to describe an imaginary experiment called 'the Chinese Room'. Imagine you do not know the Chinese language and you are in a room filled with Chinese characters. You also have a rule: you know which sign plates you have to push through the window of the room in response to the sign plates coming from the outside. People who see you doing it will probably think that you understand Chinese, while you simply pass forward the sign plates without understanding their content. This experiment demonstrates the following: the existence of syntax is not sufficient for the existence of semantics, and since consciousness works with essences, i.e., with semantics, the generation of syntactically correct but meaningless sentences will indicate that you are a bot. For example, Siri transmits your voice to the server where the neural network recognizes and processes it and then sends you the response that is synthesized into a voice. However, Siri herself does not understand what it is about at all. She is simply unable to understand it because she is only a tool using the data according to the rules written by some programmers. So basically Siri sits in some Chinese room but does not realize it. The same thing happened with the generated fragment of the of a new Harry Potter's chapter created by the AI - the machine managed to generate this text according to some algorithm it was fed with, but it had no intention to do it. It is a soulless piece of iron. 

I have a similar experience. I created a Telegram-bot, which specialized in Andrukhovych's poetry. It was trained on everything he wrote. I did not use a neural network but the N-gram model [1] to generate it. It was fun, sometimes the Telegram-bot generated something meaningful, but in general, I prefer Andrukhovych. I am glad that artificial intelligence is still not able to take creative work from us.

  1. https://en.wikipedia.org/wiki/N-gram

DSCF6390.JPG

Interviewed by Oksana Chmil
Translated from the Ukrainian by Yulia Lyubka
Photographed by Vasyl Salyga

Caitlyn Garcia