A reflection:


The poems presented in this anthology were created by an algorithm. Words from the works of writers who explore man’s quest for meaning, such as Sylvia Plath, John Milton and Hermann Hesse, are selected, categorised and placed into randomly selected and predetermined linguistic patterns to form coherent sentences. The al¬gorithm often produces pseudo-poetic nonsense, but at times offers us uncanny fragments of a curious kind of beauty; bizarre and abstract linguistic sculptures.

This anthology is a form of ‘Turing Test’, a thought experiment created by computer scientist and mathematician Alan Turing in the 1950s. The test was conceived to explore the question - can machines think? Through conversing with both a human and a machine capable of responding to questions in a human like manner, the person performing the test is asked to determine which set of responses to given questions was the work of the machine and which set of responses is the work of the human. If the evaluator chooses falsely, Turing argues, it is possible for machines to exhibit intelligence indistinguishable from humans. Whether or not you believed this anthology was written by man or machine, it poses some interesting existential questions about the evolving state of art, the human and the computer.



Many 20th Century philosophers such as Walter Benja¬min and Paul Valéry ruminated on the consequences of rapidly advancing technologies, printing techniques for example, that rendered art mechanically reproducible. Valéry argues that we must alter our understanding and treatment of art in a modern context in order to come up with new techniques and ideas. Benjamin went on to discuss art and its authenticity, its ‘aura,’ its existence in time and space at the moment it was created by a human, and how through reproduction the meaning of it is altered. If a computer algorithm can endlessly, infinitely produce writing with some poetic value, does this mean it has no ‘aura’? Or, as Valery would suggest, do we need to alter our perception of what ‘meaning’ or value is ascribed to in the modern context, in which many artists are using computers to assist in or create images and words.

The programming of randomly created acts

This idea of chance as tool in the production of art also raises some concerns about authenticity. The moments of pleasing rhythm and rhyme in some of the poems add meaning and texture to the words, but meaning that a human mind has ascribed to the words of a robot churn¬ing out randomly selected patterns of words. Does this undermine their value? If so, does this then undermine the work of surrealist ‘cut-up’ techniques, first devel¬oped in the 1920s with the Dadaists, and popularised by William S. Burroughs in the 1960s? Does meaning arise from the artist’s intention to create something beautiful, something meaningful?
The infinite monkey theorem, born out of complex theorising on the rules of probability, argues that a monkey punching away at a typewriter for an infinite amount of time could eventually produce William Shakespeare’s Hamlet. Does this version of Hamlet, produced purely through the restrictions of mathematical probability, hold and value and any other version of the text? One, for example, produced and reproduced by letterpress, laser print, pixels on a screen, the spoken word? Does the origin of the artwork matter if the words incite something within us?

Author VS. meta-author

Perhaps it is them possible to consider computational devices as artists and poets. However, it can be seen to be simply following a simple set of rules specified by the rational, conscious (meta)author –- the human, the person who wrote the code. It can’t necessarily be said that the computer is ‘thinking’ as the Turing Test explores, because it is playing by the rules I laid out before it. However the pearl at the center of this plastic and copper binary shell lies with the surprising and unexpected elements that come out of something that has an expected outcome, as some would describe as being the ghost in the machine.
The ability that the machine has to re-order symbols construct imagery in a sensical way was once considered a unique characteristic exclusive to the conscious, human brain – but no longer.



We associate the arts as being something intrinsically human. Art, poetry and literature are expressions of our human experience, a way of navigating and sharing our feelings and desires. It is becoming apparent that ma¬chines can perform many human tasks more effectively and more efficiently than us. But in what kind of state is ‘human-ness’ left if computers could also create beautiful, compelling artwork?

Software structures of the mind

This algorithm broke down poetry into its most basic components; the creativity of the author also is also the product of an excessively decadent cocktail of language, experience, memory, emotion and chance.
Simple rules of addition underpin the fundamental functioning levels of the human mind. Systems in the brain that denote whether or not synapses should pass on information in the brain to instruct the body to do certain things (move hand, wiggle thumbs, text your ex) rely on a simple ‘yes’ or ‘no’ method, a kind of biological binary that underpins biological systems. ‘Real’ human thought can be distilled down to the same patterns. It is slightly paradoxical when we consider that we may be able to create beings that are not only better at us than most jobs, but also more human. Or that we could create something that can simultaneously speak, read and write every language on earth, but also make us feel as safe and loved as a warm slice of banana bread baked by your grandma? For some, the omnipotence paradox – can an all-powerful god create a stone he cannot lift? – for some disproves the existence of a traditionally theistic god. Will artificial intelligence advance such that we can create a being that can be more human than a human being, be able to love more, feel more, create more? Does this undermine everything we thought to be fundamentally ‘human’?



Through a series of Turning tests conducted on live audiences, philosopher Oscar Schwartz provides examples in which humans write like humans, computers that write like computers, computers that write like humans and most interestingly, humans that write like computers. If we reverse the Turing Test, are some humans more human than others? Are some computers more human than some humans? Are some humans more computer than human?

Artificial intelligence as a mirror

When we ask the question, can computers create meaningful poetry? we are really asking what is poetry, what is meaning, what is it to be human? What is the category of the human? How is it different from the machine? The algorithm and the Turing test takes advantage of the fact that humans imbue groups of letters, groups of words, groups of sentences into meaningful linguistic structures. Human intelligence can’t be boiled down into a ‘yes’ or ‘no’ style test because we do not examine the world in such as simplistic way. The category of the human is unstable, shifting – what people consider to be ‘human’ or ‘robot’ poetry will be different, because our expectations of the human change with time, along with our world-view - as Valéry and Benjamin anticipated. It is this instability that makes us human, the malleable and shifting nature of our perception of the world and the space we occupy within it. Machines are programmed structures that reflect back at us the information we feed into it.
This algorithm was fed by Plath, Milton and Hesse and it reflected back to us some interesting outcomes. So when we think about artificial intelligence, we need to consider what elements of humanity we would like to feed into it for it to learn from, re-order and reinterpret. As artificial intelligence develops, what aspects of the human would we like to see mirrored back at us?