Writing with AI is Like Catching a Fish

This article is part of a series focused on how artificial intelligence is being used as a tool for writing for the screen and stage as well as how it is depicted on screen and stage, commissioned as part of MyWorld, a UKRI-funded project that explores the future of creative technology innovation by pioneering new ideas, products and processes in the West of England.

As the hype cloud of this recent AI cycle fades, the job of figuring out how to integrate new workflows into old and what this means for human creativity properly begins. In many ways, writers have most to lose from the technology widely referred to as AI and occasionally prefixed with ‘so-called’. A cursory look at the text-based applications of AI in Higher Education, publishing and particularly marketing explain writers’ hostility towards what many call the 5^th industrial revolution. Writers rightly protest AI’s wholesale adoption, citing accusation of plagiarism, legal inequities, qualities of output and erosion of income streams. While recognising these legitimate concerns, over the past year, during practice-driven research into the affordances and limitations of AI for screenwriting, I have found myself increasingly inspired to write with AI across a wide range of outputs. And I don’t mean pump it for information, book summaries or PowerPoint presentations, or to schedule my day. I mean write. The back and forth of drafting, word choice, syntax, ideation and flow. Everything that goes into making meaning in the written word.

I’ve been looking into ways to harness technology as a writing and storytelling strategy for nearly 10 years. My creative writing doctorate investigated the multimodal languages and practices of social media for creative writing as way of enabling new knowledge around writing contemporary young adult fiction and explored social media’s role in identity formation. So, when in the summer of 2023, Hollywood writers and actors’ unions went on strike to protest against the next disruptive technology to threaten livelihoods, the proximity of our relationship to machines moved a step closer and with very real consequences. Unconvinced by the corporate argument that AI could replace human writers, I devised a research project, funded by the British Academy, to investigate the increasingly blurry line between human and machine writing while providing an opportunity to revisit my filmmaking practice after a 15-year hiatus.

One of my favourite metaphors for thinking about creative ideation, which I often share with my students, comes from David lynch, who famously said, ‘Ideas are like fish. You don’t make the fish you catch the fish. Desiring an idea is like baiting a hook and lowering it into the water.’ My students and I go on to discuss how we make the conditions for ‘catching the fish’/new idea, particular places, activities that inspire us, sensory reflective practices such as journalling, watching, reading, listening etc. whatever moves you into the right state of mind for ‘catching your fish’. This process can take time. Years sometimes. But that’s the point. It’s a process. Not a destination. At least that’s not always the most interesting thing about the thing you’ve made.

So, how is writing with GenAI like catching a fish?

Image credit: ChatGPT 4.0/Sarah Gibson Yates 2025

Well, first, a disclaimer. I am not a fisherperson nor have I ever been a fisherperson, nor do I have any known affiliations with fisherpersons living or dead. My knowledge of fishing is entirely predicated on second hand observations mediated mostly through small screens (Gone Fishing, Monsters, various YouTube videos my husband, also not a fisherperson, insists on watching). But I have spent a lot of time by rivers and stared into their (often murky) depths searching for signs of life. I have also done this with many a blank page. I know there is life/words in there, it’s just sometimes hard to see it/them, let alone catch it/them, or take a trophy photo, or write them down.

This is where AI can be helpful.

AI can help you catch the fish/idea, lurking cruelly just out of sight.

This quote from American writer Joan Didion was stuck on my screen throughout my PhD: ‘I write entirely to find out what I am thinking, what I’m looking at, what I see and what it means. What I want and what I fear…’ (Didion 1976, p. 2) It is much quoted among writing circles for good reason as it seems to get to heart of what writing is about and got me through the good times and bad as I struggled to find and land what I wanted to say.

As with all writing, writing with GenAI can help you find what it is that you want to write.  And sometimes it can get you there quicker. It can also help write you out of blank page as this 2023 paper by Iona Gilburt explains.

However, ‘catching ideas’ with AI is not like dropping a line. It is a net. It can help you catch a lot of ideas. Often too many. So, switching from the image of a solitary fisher to a trawler with its oversized and indiscriminate net, the writer will soon realise most of the fish/ideas that you have caught in your (AI) net will need to be thrown back. This takes time, care and a clear sense of what you are looking for. Too big, too small, wrong species all together? These ‘wrong ’uns’ must be ignored, deleted, or re-written into refined prompts adding or subtracting according to how they fail to align, engage, disrupt, reframe or reimagine your initial prompt and according to your intentions as the 90–in this loop, writing for other humans.

So, for me, this is how writing with AI is a bit like catching a fishing when it comes to ideation, and I tend to think of all writing as a form as ideation because from first to final draft the writer is chipping away to land the idea right.

AI can help you write a path toward your idea. By crafting the right prompt and analysing the results and probably writing another 10 or so I often find myself nearer the idea I had on the edge of my vision but don’t yet see full screen. This iterative process, guided by a set of criteria defined by your individual writerly aims and objectives, can be worked any number of ways, depending on what you are writing, how you write, your position as a unique human in a particular body, with a specific []=p-pp set of experiences, in a particular place and time. These mental and physical criteria can incorporate all the myths and science around writing: ‘the muse’, subconscious babblings, the divine, serendipity, improvisation and ideas stolen/adapted from others, and run through your own creative algorithm just as artists have been doing since all time.

For writers used to getting hired and fired within the production systems of Hollywood, the suggestion that AI tools might cheaply churn out scripts for producers wishing to duplicate the story formulas of past successes, felt like a death toll. The widespread belief that a computational system trained on pre-existing data and programmed to recognise story patterns could replace a skilled human with life experience and a nuanced understanding of the complex context in which they write was a B2B conversation. Corporate America talking to itself about how to maximise profit and minimize financial risk while appearing to offer transformative benefits to the workforce in the same breath. Artists were right to ring a clear loud bell of caution, not to mention pessimism as the short-term implications looked harrowing and reflected a wider, longer held belief that threatened to reinforce a historical and persistent devaluation of the writer, or, perhaps more accurately, the processes of writing. Revealing our capitalist economy’s emphasis on product, not process.

Screenwriting is not alone in having to confront these issues. We see them being confronted across many creative disciplines but while other practices and industries have their own issues, the problems for (screen)writing, in this peak of text-based AI and in the face of widespread misbeliefs about what any creative writing actually is, feels particularly pertinent.

‘AI, while impressive, only really excels at a very specific type of writing…and …is only better than people with no particular expertise in writing’ (Lynda Clark, 2022, p. 135).

This is where the writers come in. If we believe, as most writers do, that writing is more than just the accumulation of words on a page, that it is ideological, reflecting and shaping beliefs, systems and values. We must therefore work with AI to show the world that our skills as writers, which lay in staying attentive to how words show our world back to us, are still very much needed in a post truth world saturated with populist rhetoric and misinformation.

In a recent research project I explored creative human machine collaboration in screenwriting and filmmaking within a new context of care that emerged in response to the #metoo movement and climate crisis. In it I built on Clarke’s notion of creativity amplification (Clark 2022) and developed a framework with which to net the right kind of fish/ideas (or at least increase the odds of doing so). I describe The Screenwriting with AI Framework (SWAIF) in more depth elsewhere (Gibson Yates 2025), but to summarise, SWAIF emerged from an initial messy process of devising prompts based on my research questions and aesthetic intentions for the work, assessing the results according to a range of specified criteria, intuition and understanding of context in which the work would sit – and budget (I needed to produce this film after all). It provided a structure to my workflow while preserving the collaborative approach to fostering creative spontaneity and flexibility I was looking for; to enable authentic moments of human-machine intra-action and entangled meaning making.

Importantly, it offered a map to stay on course. It’s easy to lose direction once you start creative writing with AI. You soon find yourself with a large body of ideas, scene draft variation, character profile, different plots, reams of dialogue exchanges all offering different and seemingly viable variations on your story. SWAIF helped me stay on track, serving my research question to explore the intimate and affective spaces created when human and machines intra-act, but also allowing me a generous amount of creative freedom.

Writing with AI will not be for everyone and resistance will no doubt continue, but soon, as AI becomes the default component in all software, the time when it could be hard to write anything (with a digital device) without it soon. Each new wave of tech has its resistors and rightly so, but eventually the evolutionary mantra of adapt or die comes to pass. Reflecting on the history of writing and literature’s historical resistance to technology, from the reluctance to embrace the Word Processor, to debates about whether contemporary media should be referenced in ‘serious’ fiction – from TV in the 80s to the internet in the noughties, I am reminded of an influential essay by the American author David Foster Wallace. In a much cited speech to graduates at Kenyon University, Ohio, in 2005, curiously also featuring fish, he gave this advice on the importance of engaging/staying attentive to the stuff of everyday life that surrounds us:

‘There are these two young fish swimming along, and they happen to meet an older fish swimming the other way, who nods at them and says, ‘Morning, boys, how’s the water?’ And the two young fish swim on for a bit, and then eventually one of them looks over at the other and goes, ‘What the hell is water?’’

This story works on many levels. For myself as a writer observing and trying to work out the best way through the world for myself, for loved ones, for fellow writers, makers, teachers, and for generations who come after us, the idea that we should pay close attention to the stuff we barely notice among the daily stresses and strains of life is important. It many ways it fuels my creative focus and perhaps goes someway to explain why I, for one, will continue to write with AI.

‘Technology enables us to change our vision, and it allows us to change what we do with that vision; where we look, what we see, and how we act as a result’ (James Bridle, 2022, p.138).

AI is already the invisible stuff all around us. Understanding how to work with it, and to some extent the mechanisms and systems that underpin it, can only benefit us. By increasing our knowledge and understanding of the possibilities and limitations of writing with AI we can grow a better understanding of how we can intra-act proactively, creatively, and even with hope.

‘I am sure, that our future as Homo sapiens is a merged future with the AI we are creating. Transhumanism will be the new mixed race’ (Jeanette Winterson, 2021, p.262).

I see working with AI as more than simply a process of figuring out how to use the next tech tool, but rather, as physicist Karen Barad terms, a way to intra-act with it and the other complex life systems via AI applications in life sciences, conservation, climate intervention, all of which are already helping us see anew (and act anew). It is a way to speak directly to our entangled relationship with technology and an opportunity to gain new meaning and insight into the complex interactions of digital and material agents and how they can be harnessed to benefit all of us. Writing with AI is a space for holding ideas about the potential of these posthuman futures, a way to hold them in a space driven by community, care and love.

Acknowledgements:

The research project underpinning the development of this framework, Creative Human-Machine Collaboration: screenwriting a short film screenplay with generative AI and its production within a context of care, was funded by The British Academy Talent Development Award for Early Career Researchers and supported by Anglia Ruskin University. The work was carried out in 2024. Find more information, including the screenplay and watch the film here: www.sarahgibsonyates.net

References/Suggested Reading:

A Little of the Heart (2025) Directed by Sarah Gibson Yates. https://vimeo.com/988573902

Barad, K. (2007) Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning. Durham: Duke University Press.

Bridle, J. (2022) Ways of Being. Animals, Plants Machines: The search for a planetary intelligence. Penguin.

Clark, Lynda. (2022) ‘Towards “Creativity Amplification”: Or, AI for writers, or beating the system’. Creative Writing in Practice, pp. 134-145, https://www.nawe.co.uk/DB/current-wip-edition-2/articles/x11-towards-creativity-amplification-by-lynda-clark.html

Didion, J. (1976) Why I write. The New York Book Review, vol.22, pp.270.

Foster Wallace, D. (2008) https://www.theguardian.com/books/2008/sep/20/fiction

Gibson Yates, S. (2025) Creative Human-AI Collaborations in Screenwriting: co-writing a short realist screenplay with generative AI. International Journal of Screenwriting Special Issue on AI issue. Intellect.

Gilburt, I. (2024), A machine in the loop: the peculiar intervention of artificial intelligence in writer’s block. New Writing, Vol. 21. No 1. 26-37.

Heti, S. (2023) According to Alice. https://www.newyorker.com/magazine/2023/11/20/according-to-alice-fiction-sheila-heti

Ishiguro, K. (2021) Klara and The Sun. Faber.

Lynch. D. (2009-2025) various video interviews including David Lynch’s Secret to Catching ideas like fish, https://youtu.be/vT3Znxzp4h8?si=zM3AWTIy7-hnlNe4, and, David Lynch: Where do ideas come from? https://www.youtube.com/watch?v=Fxr-7O1Bfxg

Silva, H. (2023) My Child, the Algorithm. Footnote.

Winterson, J. (2021) 12 Bytes. How Artificial Intelligence will change the way we live and love. Vintage.