User:Alessia/Writing
Poetry as a human experience
I remember writing my first poem in primary school. I was crawling in my childhood discomfort, after a particularly bad argument with my mother, one that I don’t even remember the details of. I sheltered myself in a corner, pressing myself against the walls. Without shedding a tear I grabbed the first surface I could find, funnily enough a catechism book, and on the first page I poured out the most intense poem of my life.
I don’t recall it at all, nor I found that book again.
That was the beginning of a very common pattern for me: the inability for me to not be able to open myself emotionally. At the same time it was the start of something new, a new way to seek light around compressed locked up feelings, a way to let emotions flow without making me die.
I don’t feel that was the real answer to many personal problems, but indeed writing helped me, it was a useful tool.
Not through writing prose or diary clear confessions of feelings, but with poetic language, a confused dense symbolic one, that became the truest form of visualising feelings I couldn’t understand, let alone heal from.
As I continue wandering through reality I’ve become more fascinated by poetry, not just as a personal outlet but as a medium. I’m intrigued by how others use it, what it means to them to write and experience it. All this, while I still need to decipher poetry as a medium for myself, what it means to me.
There won’t be any point of this research that aims to answer the question “what is poetry?”, I am firmly convinced there is no definitive nor useful answer to that, I will instead explore the many ways poetry can be understood, created and played with. If there is a point of truth I believe in it is this: poetry is a human experience.
Poetry as a political experience
https://www.yahoo.com/news/something-doesnt-love-wall-poetry-matter-age-trump-090027168.html?soc_src=mail%26soc_trk=ma
In the political world, poetry is nothing new, as well as other many art forms.
Art permeates society, culture, and daily life.
The increasing accessibility and democratisation of art, once reserved for the elites, have also made poetry more available to the general population.
Robert Pinsky, in (add) , said about poetry that it is “essential to the whole democratic project, It was traditionally part of the education of ruling classes, and in our country the ruling class is supposed to be the people, everybody.”. Even if Pinsky is referring to the US poetry scene I would like to extend these same words to most of western experience of poetry, in contemporary times.
With this broader accessibility, poetry gains the potential to act as a tool for social change and political action. For example, initiatives like those promoted by organizations such as Poetry Therapy highlight how poetry can inspire reflection, healing, and activism. This way, poetry continues to play a vital role in the civic engagement on addressing society’s issues.
What is interesting is …
This is what, for example, the The Andrea Mitchell Center for the Study of Democracy of the University of Pennsylvania is doing: exploring the interrelation between poetry and democratic politics.
https://amc.sas.upenn.edu/poetry-and-politics
(Francisco's interview about poetry as a social justice and fighting tool)
Poetry as a feminist experience
- Feminist poetry as a response to exclusion from male dominated literary spaces (costellazione parallela)
- Historical connection between poetry and activism in women civil rights movements
- How does poetry address issues like gender inequality, identity, intersectionality, how it provide a space for reclaiming silenced narratives
- Critiquing gender roles and celebration of nonconformity
Chimamanda Ngozi Adichie (Nigeria), Kamala Das (India), or Forough Farrokhzad (Iran)
https://en.wikipedia.org/wiki/The_Dream_of_a_Common_Language
current situation of women’s poetry in English and critical attention to your pioneering contribution to the field of feminist literary criticism
Poetry as a therapeutic experience
https://poetrytherapy.org/
https://poetrytherapy.eu/
Can poetry help our mental health?
Art has always been a way to soothe pain and express the inexpressible.
Few studies highlight the wonderful property of poetry reading and listening as a tool to improve pain levels, give more hope, increase cognitive functions, and improve capacity to cope with stress.
https://www.liebertpub.com/doi/10.1089/jpm.2015.0528
https://www.jstage.jst.go.jp/article/tjem/235/1/235_17/_article
Reading poetry can help our wellbeing, soothing the experience with difficult feelings and with things that fall out of our own control. We can feel less alone, reminding ourselves there are others out there feeling the same as us. Learning about loneliness in others can let us feel less lonely, at the same time the intimacy that poetry is built around gives people a nest, a place for reflection in their own mind that can foster understanding and compassion.
When it comes to youngsters, children, we can see how poetry has an effect on the emotional wellbeing of those that are challenged by intense hospitalisation, stress and anxiety.
https://publications.aap.org/hospitalpediatrics/article/11/3/263/26172/Effects-of-a-Poetry-Intervention-on-Emotional?autologincheck=redirected
While creative writing and narrative medicine have been widely researched in health disciplines, poetry therapy remains an emerging area of study. Few studies exist on the integration of poetry therapy into psychiatry and psychotherapy, though pioneering work by researchers like Heimes (2011), Mazza (2003), and McCulliss (2011) has laid a nice foundation for further researches.
Mazza’s RES multidimensional model, for example, is one of the tools created thanks to the interest of researchers on poetry, that aim to frame poetry therapy within a clinical environment.
This so called model function following three components:
- Receptive/prescriptive Mode: reading or listening to poetry to evoke therapeutic responses
- Expressive/creative Mode: encouraging individuals to write poetry to express emotions
- Symbolic/ceremonial Mode: using poetry for reflection and ceremonial purposes, offering a sense of structure and closure
( Putting stress into words: Health, linguistic, and therapeutic implications☆James W. Pennebaker https://www-sciencedirect com.hr.idm.oclc.org/science/article/abs/pii/0005796793901054)
While poetry has been called for its therapeutic benefits, it’s important to approach this anyway critically. The suggestion that poetry inherently makes people better, like any other form of creative art, is problematic and I don’t agree with it at all.
Poetry doesn’t make people feel better.
Poetry is a tool, and like any other tool, it can be used constructively or destructively.
History offers examples of poets who lent their voices to oppressive ideologies and tyrannical regimes. In this sense, poetry is not inherently an “enlightenment tool”.
https://aliceosborn.com/why-is-poetry-important-to-our-world-today/
The first poet that comes to mind is Gabriele D’Annunzio, one clear cult figure in italian culture, I love and praise his own depiction of nature, he is still one of the first poet I learned to appreciate and love, still he was heavily influenced by the early italian fascism, his own rhetoric glorified violence and militarism, his own words were used by Mussolini.
D’Annunzio led the occupation of Fiume (modern-day Rijeka, Croatia) in 1919, blending nationalist fervor with theatricality, incorporating poetic symbolism and language to justify imperialistic violence
What about Ezra Pound? Alfred Rosenberg?
What about the full distortion of words of poetry?
The nature of poetry helps its distortion, poetry is not made to be clear, the fact that propaganda turns words to its own desire is not something new, not even just with poetry, it is a common fact. I could go hours just finding more examples of personalities whose words were distorted by propaganda or they worked to enhance the propaganda itself.
This just highlights the fact that poetry is indeed a neutral medium that reflects the intent of its creator, as any medium out there.
Poetry as an artificial experience
An example
I always believed, and already wrote in this compulsive essaying, that poetry is a human experience. That’s why I’m writing this chapter, because I am trying to find a way to disprove this idea of mine.
Last year, while volunteering at Poetry International in Rotterdam, I encountered a project that stuck with me. It was intriguing while being unsettling.
The project was a photo booth, but instead of producing mere photographs, it generated AI poems based on the images it was capturing.
I got mesmerised by the sublime action of the machine. What was the point?
The whole project had a strange sublime aura. The project felt a bit like an attempt to rebrand poetry as something that could survive the AI age.
The funniest part was that you had to read the poem extremely fast and make a photo even faster, in a bit it was gone and the qr code of the poem vanished.
It felt uneasy on how come this was an idea, to be put in a festival that was focusing on human made poetry. Not just an interactive game that showcased the AI capabilities but a full art piece that aimed to bring poetry closer to everyone.
https://www.vouw.com/poem-booth
The poem booth, the AI Poetry machine that turns portraits into poems.
In this interactive art installation there was one point I don’t enjoy about the contemporary art scene: the obsession with interactivity, the tendency to prioritise interaction for its own sake.
Such projects always in their aim to be consumable reduce art to a mere product, something to be tokenised, an artifact of our consumerist age. So, as I said I don’t like it, but that doesn’t mean anything, as my higher belief is that art is a mirror of our society, so this photo booth fall into that, a sad mirror of our society.
In this project interactivity as a concept plays an important role.
When you think about interactivity in art installations, it usually means that your actions directly influence the work, changing its behaviour or creating some kind of unique output. But does that really happen in this case with the poem booth?
To me, it doesn’t seem so. The act of taking a photo to generate a poem feels like little more than an excuse, a way to give participants the illusion of agency, while chat gpt will produce a random poem for you in a way or another. The interactivity feels superficial, the poetry output is too.
https://www.vouw.com/post/how-were-using-chat-gpt-generative-ai-to-bring-poetry-closer-to-everyone
While digging a bit into the story of the project I came across a post describing its showcase to children. In the key takeaways of the event this is shared: Kids loved it. A few quotes: "I'm addicted," "I want one at home," and "I could stand here for hours.” Reading this made me shiver, knowing many ipad kids will say the same thing for any digital device that will play on their (and our) serotonin scrolling fast content addiction.
The last point shared is: Parents took pictures of the screen to keep the poems. This made me even itchier. This is clearly a product masked as an art piece installation, and that’s even more clear when the developer directly told us they could explore two directions: creating a Poem Booth for (semi) public spaces or developing or a Poem Booth specifically for kids, which we envision being placed in waiting rooms, schools, and libraries. The first could be even fine for me, the second sounds just dystopic.
You can’t just drop a tool like this into children’s spaces without considering the ethical implications. There’s a cognitive dissonance here: imagine a child learning that poetry is something deeply personal, something that takes time, introspection, and heart to create, only to then encounter the Poem Booth and see its output described as “poetry”.
Even if that’s just sad alone.
But more than that, we already live in a world that constantly pushes us to move faster, to work harder, to produce endlessly, we already normalise it.
One of the values of art, and poetry, lies in its ability to slow us down, to help us step outside of our cycle.
Art is still something that can free and soothe misery, poetry does that too. If we suggest to kids that poetry is fast then the whole point of art as a human creation won’t subsist because humans will surely be slower. A kid that writes one poem per day, and then finds out that the machine makes thousands of poems for hours without grammar mistakes and that is easy for everyone to understand, won’t be encouraged to keep going.
I understand the point of many observers (to add) that automatisation could make human labour disappear while enhancing human creativity. I feel this might anyway be a slippery path.
There are machines that bake perfect cakes, mass produced, ready for sale. Yet people still bake cakes at home, and there are even bakeries selling artisanal cakes for premium prices.
This doesn’t prove automation democratises creativity, this to me speaks loud about privilege.
And that’s the crux of my argument. AI won’t democratize art or poetry. Instead, it will deepen existing differences, highlighting the privileges of those who can afford the luxury of human made art.While others will be left with the automated version, unable to choose otherwise.
A project that is quite similar to the poem booth is the Poetry Camera, by two designers, Kelin Zhang and Ryan Mather. The premises are quite similar, the technology involved is the same, the concept of the work is quite different tho. We see the Poem Booth selling itself, advertising itself as an educational tool, while the Poetry camera is much an art experiment. As stated in their website the Poetry Camera is an art project, a celebration of creativity, community, and play.
(https://github.com/carolynz/poetry-camera-rpi)
- The Poem Booth in action
- The Poetry Camera, with one of the output
Poetry feels as a refuge for many, a safe island in a sea of consumption driven madness. Could AI and the multitude of art projects similar to this destroy that safe refuge?
I often think about the many artists’ initiatives that tried to ban AI tools and AI media sales on popular art focused social platforms.
While I understand their intentions, I can’t help but feel that that effort was quite useless.
AI will take a piece of the cake, of the market, in a way or another, and independent artists and freelancers will be the first to fall.
There’s a pressing question then: how will AI reshape, remodel, our conception of art? How will it enhance or destroy the intrinsic cultural value we assign to it?
The value we give to art is relatively new. Before the Renaissance I would say the artist wasn’t as much considered at all.
Art was considered an artisan’s craft, a functional and communal activity more than an individualistic one. Artists, as we understand them now, didn’t exist at all. Artisans were skilled labourers fulfilling commissions for patrons and communities. They didn’t have any concepts like copyright or ownership.
The figure of the artist as we now know it, the creative genius emerged thanks to renaissance humanism and the later social and cultural transformations. Artists became cult figures, revered as gods, their work seen as unique expressions of their inner world. The concept of copyright popped up later, when art became something that could be made on a larger scale, with printmaking, and people were copying quite a lot and gaining too much money, so Durer didn’t like it and for the first time copyright was claimed as a right.
Today anyone can become an artist, the boundaries of what is art and not changed immensely. This democratisation of art has led to amazing examples of creativity and few arguments too.
If everyone is an artist, and everything can be art, then the AI can then become an artist too, and its art can be valued as human art is (or at least the good branded human art). That’s just logical, isn’t it?
Still we may have much resentment toward this new face of technology, as we position ourselves much at the center, we value ourselves on top of the food chain and for this reason we cannot find a way to accept how something we created as gods will make something that resembles our making. I always think about how much anthropocentrism still plays a role in all this, a lot is my guess.
This tension is fascinating.
Could this shift in the art and creative world push us toward a broader understanding of creativity? Or will it make our resistance stronger toward changes that don’t put the human artist at the center?
AI is becoming increasingly a sparkling and precious object of attention.
I had a nice conversation with a colleague of mine that fully perceived AI as the only way to go, the only useful path in the work industry. You know they need people to work on those things, anyone is now participating in the race to become an AI expert. I want that shiny, polished “AI expert” glow on my linkedin page too, I want it! I want it!
Narrative and AI
My conception of artificial intelligence was limited, and still is.
In a world like ours, this is no longer permissible, one is called to know more, of a technology at the very core of the public debate, a technology that overtakes everyone and everything and claustrophobically becomes a political, social and economic object.
Since not knowing enough about AI, as just noted, is no longer possible, unless one wishes to live in remote isolation (and even then I am not sure that at some point Open AI will not come knocking at the door), here is a perfect way to delve into the topic: making it
a chapter of my thesis.
Artificial intelligence has grown fast enough to be scared, profoundly influencing our political, social and economic circles. Its impact on poetry is illuminating.
Poetry is often seen as a deeply intimate form of expression. Historically the value of the medium has been tied to the poet’s intent, experience and craft. When the author spot is left vacant few complications could emerge, not just for poetry but all forms of communication.
One of the main focuses of skepticism over AI is authorship.
People survive thanks to creative jobs, thanks to writing too, the disappearance of the author will inevitably bring many capitalistic problems, machines stealing our jobs.
Who’s going to be blamed? Words are razor, even in an overwhelming content filled world like ours authorship is still important to give responsibility to people for their actions. Writing a full hate speech, that could aim to fuel hate online, move the masses, hurt somebody, produce the content just described with AI, dashing the rock and hiding the hand, will bring many problems. AI could then become a tool to hide authorship for the good and bad.
Poetry is a dialogue between the reader and the poet’s personal perspective, AI generated poetry challenges this vision of it by eliminating the author’s intent. AI creates poems by identifying patterns and statistical probabilities in language. Many argue that this lack of human intention and personal connection may produce aesthetically pleasing patterns yes, but void verses.
Others counter that the "death of the author" principle, where the meaning of the art is shaped by the viewer or reader, not the creator, validated AI poetry too, as equally meaningful. Sometimes, knowing the context of why a poem is written changes how you feel or think about it.
For my own self I can say I don’t feel human poetry and artificially generated poetry should be put on two different levels intrinsically, as I believe the AI poetry is still created thanks to an impressive amount of human generated data. What I feel is wrong is the production of it alone, the impact of AI to our world and existence, and the fact that humans are getting money for content generated from other people's creativity.
And, in a world where people need to get money to get food, is in fact exploitation.
When I see a lot of discussions around AI generated content, what is visible is those that support AI are generally the one that won’t be influenced by its taking on the market (for now), the one that don’t value the fact people would like to survive with their writing, drawings, creative production and now their might not anymore. Those are usually people that have already a prejudice against the creative arts jobs, those that would ask an illustrator why they don’t find a real job.
The concept of Art is redefined constantly by human evolution, social and political changes, Art is not an unscratchable monolith, it mirrors societies. I would never say AI art is not art. In a broader sense, it can be enjoyable or not. I am quite sure Warhol would have loved AI, he would have still taken advantage of it and rebranded it, put his name on it, then sell as much as possible of everything.
Great poetry is what you feel is great, there is no objective metric to define what is art and what is not art, this applies to poetry too. There is poetry that is great to a great amount of people, but there is no universally great poetry.
I think it's because what people really want from poetry is feeling a connection. By design, AI is better at that in a vacuum. It can create things that are really close to "average," so overall people will understand it better. But it can't really get personal to build a stronger connection past that”.
By design AI is great at creating “average” content, and poetry. It is not going to create new types of poetry out of nowhere. This “averageness”, this flattening, can be a double edge sword, as AI is built to create easy answers as much as possible. All poems could then become dull, using the same language and expressions, making poetry predictable. At the same time it could help people not used to consuming poetry to enjoy it more, as it is more easily digestible, or that’s the common claim.
I might argue with this last point because there is indeed quite some “easily digestible” poetry out there already, poetry that became famous exactly because it was “easy” and thanks to that it reached a vast audience.
There are many poets working on instapoetry already.
For example, Rupi Kaur. Her instapoetry is emotionally resonant with the masses, and this made her one of the most famous poet out there.
Is that right? I don’t enjoy Rupi Kaur poetry, that’s a personal stance, I feel I am not even part of the audience that the poetry is aimed for, but still I can see how the simplicity of her stanzas was immediately relatable to contemporary readers, and it probably created interesting thoughts in the audience, and that is what is important.
She's quickly and easily comprehensible, also forgettable, following our daily content that is disposable and forgetful quite as much.
While AI generated poetry can mimic and will probably soon mimic perfectly language and themes of human written production, skeptics warn about the perceived loss of depth.
This reminds me a lot of the allure that goes around traditional printmaking.
While I was studying at the fine art academy of Venice I got familiar, on a much surface level, with the elitism and desire to close the box of the whole printmaking craftsmanship in a sort of protectionism that wasn’t helping anyone, if not the main professors and characters that dotted that environment, full of their insecurity wanted to be raised as gods.
A glossy, empty world. This has led to much of the traditional graphics disappearing, a lack of desire for openness hidden by an insatiable desire for tradition and good old days.
Dusty altarpieces.
What traditional graphic and printmaking tried to do, in many countries, was rebranding it as high luxury art production. In a way it worked, exactly how all the other art production at some point from artisan became artistic, and with it its status changed and the artist became important, while it wasn’t before.
Now what I see is a new turn on the value of authorship that is much interesting to explore, on a mere curiosity level, not so fun to experience as a wannabe poet/printmaker.
Bad luck.
Poetry is a cosmos of meaning, is navigating that cosmos
Technique and craft are just the tools. Aesthetics are not absolute. The core of art is connection
We still talk about AI as a fully conscious entity, like a creature.
Popular media often anthropomorphised AI, this depiction overstate the autonomy and intelligence of the technology that is still far removed from the current AI narrow capabilities.
AI is as well used as a political tool to enhance public anxiety, while still taking out ethical concerns the public overly misunderstand the concept of AI, giving all the faults to the tool while it’s indeed the hand that uses the tool wrongly that could make life worse for all of us. The limitations of a prevalent fictional narrative is anyway part of human nature. We see how thinking and dreaming about intelligent machines is ancient, and with dreaming here that narratives come to play, as the way to develop people’s engagements with new knowledge and their applications.
Narrative is a fundamental part of human development, disruptive technologies have sparked intense debates, we can see that with nuclear energy, genetic modification, and AI too. Such narratives give out a rich source of imaginative thinking about artificial intelligence in relation to explorations like immortality, consciousness, value alignment, governance.
There is a strong tendency in fictional narratives, and then in real life, of intelligent machines to take humanoid form. This anthropomorphisation of AI can be understood in many different ways.
First, the belief, already cited in the paragraph above, that humans represent the pinnacle of intelligence. This anthropocentrism, quite widespread in the West, shapes how we imagine other intelligent beings, and how we value intelligence as an inherently human value.
Humanoid machines, and artificial intelligences are often conceived as performers of human labours, as slaves, it can be perceived even in ancient times with Hepaestus’s attendants made of fold, C3PO in Star Wars let’s say. The easiest way to represent such machines is metallic extensions of the people they are mimicking.
Visual storytelling was fundamental for the popularisation of such pre concepts describing artificial intelligence, as humans rely on bodies, and storytelling itself tends to privilege human description, or humanoid representations. It’s natural then for intelligence machines in these stories to take a human form too.
This tendency anyway has consequences.
An example: AI becoming gendered, being associated with sexual characteristics, even hyper-sexualised. These ways of portrayal shape not only how we imagine AI but also how we imagine the future.
As Friedrich Kittler wrote in Gramophone, Film, Typewriter, “Media determines our situation.” The media sets the terms by which we live, socialise, communicate, organise (Pattern Discrimination, introduction), and this is why I would like to understand the truths and falsehoods surrounding AI, it shapes not just the discourse but our collective imagination of what AI can and will be.
The roots
Even if AI feels like a recent technological development, the groundwork for AI began during the 1900s.
While the first half of the century was much more focused on creating an artificial human being, driven by fictions, from Čapek creating of the concept of a “robot” (robota, czech for work) to Nishimura building the first robot in Japan, the second half was the time when researchers and artists started thinking about intelligence as a concept, how to create a machine that think. Thinkers always struggled to define intelligence, every definition of it comes with its cultural bias, it’s a concept that is shifting while new technology’s advances are made.
The term AI was created in 1955 by John McCarthy who defined it as “the science and engineering of making intelligent machines”. From that moment AI captured the public imagination, with films and books turning it into a mainstream concept.
The shift of interest from physical humanoid machines to intangible intelligent softwares, programs, and chatterbots, was something that happened in the second half of the century.
This did help fictional depictions of AI, and the increasing popularity of it, but didn’t help the research itself, as the media narrative led to sky high expectations, people were unimpressed when AI failed the promises of fiction.
Despite the lack of funding from the 90’ on incredible advancements were made, like the Roomba (that still gets stuck on carpets to these days) and new wonderful commercially available speech recognition softwares, track body movement systems, Siri as the first popular virtual assistant. While these developments didn’t follow the grandiosity of fictional AI promises, they easily integrated into everyday life.
When it comes to AI that writes, we need to take a step back.
The idea of using algorithms for creative writing, so for poetry too, goes back to the rise of computers. In 1677 John Peter published a treatise titled “Artificial Versifying”. With its algorithmic approach it was one of the first examples of using an automated hierarchical set of instructions to create well formed lyrical lines.
In 1845 John Clark built the first machine that implemented these principles. At a rate of one minute per verse the machine was able to compose amazing latin hexameters.
(https://slate.com/technology/2017/08/behold-the-amazing-poetry-generating-machine.html)
From that moment many machines were built and programs created, all designed to write poems, aiming for great flexibility and sophistication.
Chatgpt and PaLM stand out in the present day because they were not designed to produce poetry at all, still they can. They are both "self learning" AI. GPT for example is a predictive engine that comes up with continuations after being fed with data. The way these self learning AI are responding to our input is through language, but obviously its actual basis is always numerical. This kind of architecture is called "transformer", or "generative-pre-trained transformer".
The 90s were the times when the creative attribute started to be questioned to AI. Margaret Boden, cognitive and computer scientist, and Douglas Hofstadter argued against each other about the AI's creative capacities.
There is now an important confusion and arguing in literary and art criticism, after having repudiated the concept of aesthetic quality, here it is reintroduced to analyse AI generative art, after decades when the contemporary art scene disowned it.
That brings us to the present day.
We are not in the AI winter anymore, the research on AI funds is fueled by any government and corporations on the planet.
Politicians and tycoons are smelling the sweet perfume of profit when they see AI applied to War. As Peter Nowak argues, technological innovation is often driven by industries catering to our most basic instincts: military, sex, and food.
I might argue with that, knowing that fast food is not really in my personal basic instincts and I sense some US egocentrism here, anyway our world is still western centered as a matter of fact, so I don’t like it but still I can agree on this view.
Actually, I would like to change his stance a bit, saying that : technological innovation is driven by the translation of basic instincts into the interests of the ruling class.
The truth is, we can never fully know what is happening behind the scenes. AI is already being used as a tool of war, despite public facing gestures like Elon Musk and others signing open letters calling for a ban on autonomous weapons. The trajectory is quite clear: AI will be developed for war, for sex, and for food.
I will use AI for poetry, arguing that Poetry is part of the basic instinct of the human being.
It is the medium through which we express our most profound emotions, our sense of beauty, and our place in the universe. AI could shape poetry in ways we haven’t yet imagined, expanding its communal and exploratory potential.
Instead of succumbing to the consumerist, hyper speed narrative of machine generated poetry, we should use AI to slow us down, to reconnect us with the human soul behind art. It’s not about replicating poetry but amplifying its reach, offering new ways to engage with language, emotion, and meaning.
What I would like to work on is how AI could enhance (or not) poetry.
But how do I design something like that without falling into the same traps I’ve been critical of? How could I use the same technology in a way that feels less consumerist and disposable? Is it even possible to step away from the tokenized, attention-grabbing nature of so many AI art experiments?
Can I find a way to not rely on GPT? Maybe an installation that feeds itself?
Maybe I should create an installation too but with another type of output, something that won’t feel so sublime and just created for the sake of it, that actually shows the way I see poetry, but wouldn’t that make the installation more about me than about poetry itself?
Maybe AI in poetry is indeed an embodiment of a communal experience, to generate poetry that reflects a collective voice rather than an individual one?
Many questions.
AI researches and Art projects
Wandering around I found many interesting projects using AI models, beyond chatgpt, to create poetry or that focused on poetic expression.
An amazing project was conducted in the Cornell Humanities Lab, focusing on poetry and artificial intelligence. Morten Christiansen, Laurent Dubreuil, Pablo Contreras Kallens and Jacob Matthews worked on this project that explored the interaction between human and AI generated poetry.
https://blogs.cornell.edu/humanitieslab/poetry-artificial-intelligence/
https://harpers.org/archive/2024/07/metal-machine-music/
The project started with a poetry Turing test, so an experiment between human written poetry making and the machine one. This test was used to search for a gap, or lack of the gap, between the performance of a computer and that of a human.
Making chatgpt 3 work on poetry was quite disappointing, surely at the beginning, for these researches, as GPT-3 didn't recognise sounds. It could understand repetitions and end of lines rhymes, but just that. When InstructGPT was released improvements were fine tunings were made in the model, enough to fix some stereotypes, prejudices, and biases that so commonly surfaced before. InstructGPT is an example of how the "self-reinforced-learning" activity of the model is not enough, corrections and fine tunings are needed by human agents to identify bugs, most importantly bad content.
This new found political correctness might have helped how the ai utilised rhymes and meters, maybe because a closer on the users' prompt was made. One thing that emerged was a new "non-determinism" attribute, the fact that every output is different from one another, or at least try to be so.
(Turing seminar 1947 "Lecture on the automatic computing engine")
During the experiments the researchers found out about the first knot. The AI is a self learning program, still it works with a collage technique. The AI is memorising pieces of text and play tangram with them, glueing those pieces with connective tissues of synonyms. But, isn't human writing too thanks to a collage of influences that we can’t still map out completely? About this the rhetorical question “Who can be certain that ‘original work’ that he has done was not simply... the effect of following well-known general principles?”
The study highlighted how most of the time human poetry and AI generated poetry was difficult to distinguish. There are few reasons for it. The participants were displaying biases, underestimating human work and overestimating AI’s capabilities. AI strength lies in imitation and consistency. The study reaffirms that both humans and AI can produce uninspired work. The real challenge that still comes answered is understanding why certain works resonate more than others in a personal way.
This research is quite enlightening considering the broader debate it ties in linguistics.
When AI demonstrates a great technical proficiency it then lacks true originality, but what is originality then?
Over the last years thanks to the multitude of copyright claims after the boom of OpenAI, the main AI companies have advertised new positions for writers and poets. OpenAI organised a deal with Axel Springer, an european multinational media company focusing on printing and publishing (https://www.axelspringer.com/en/ax-press-release/axel-springer-and-openai-partner-to-deepen-beneficial-use-of-ai-in-journalism), for access to its content.
The AI companies know how much AI needs new fresh content, surely while the content that is freely online is already being written with the help of AI tools.
A dystopian horrible future seems to be a way this system will evolve into, like other industries, the AI will be fed by human improvements by low wage employees, contributing to let the AI find that “originality” that comes with just being alive.
This seems like a bland future, something Turing could be right about.
Another project that got quite famous is the Verse by Verse by Google.
https://sites.research.google/versebyverse/
(https://www.engadget.com/googles-ai-poetry-verse-by-verse-202105834.html)
It is a project that creates poetry inspired by classic American poets.
You can select up to three poets for inspiration, you can pick the structure of the poem and the tool will compose the lines for you.
Other project I want to explore:
https://artsandculture.google.com/project/poetry
https://rossgoodwin.com/
https://artsexperiments.withgoogle.com/poemportraits
Madeline: Poetry and Art of an Artificial Intelligence
https://www.mdpi.com/2076-0752/11/5/83
https://blog.on-page.ai/ai-and-poetry-writing/
https://www.nvidia.com/en-us/research/ai-art-gallery/artists/allison-parrish/
https://www.scienceopen.com/document_file/add6b15b-66ea-4bc5-8368-6c537204f277/ScienceOpen/329_Mills.pdf
https://en.wikipedia.org/wiki/Sunspring