Freedom to be, to develop, to live, to create. Creative risk is a set of affirmative choices – artistic or political or both – made in the service of positive freedom and in the hope of flourishing. Creative risk is, above all, the feeling of making those choices. Risk is a state we experience.
Humanities scholars are constantly scolded by technologists for seeking the one thing that machines cannot do as well as humans. But we need to be clear: machines and AI systems don’t experience creative risk. Nor do they enjoy or suffer its consequences. Organisations, including universities, governments and corporations, manage risk – which they define in reputational, legal and financial terms – but they, too, don’t experience it as our human bodies do. Corporate definitions of risk, which aim to produce freedom from this or that danger, often hide in plain sight a worldview that is antithetical to creative risk, and positive freedom, and the human lives that host these unruly, unpredictable impulses. Timothy Snyder calls the subjective experience of our own, unpredictable, corporeal life our Leib. He writes about AI, ‘As we translate the experience of life into machine language, we risk our minds becoming something like satellite servers, to make things easier for the computers – which do not and cannot care.’ Creative risk today involves navigating the incursions of AI on the one hand, and organisational risk management, on the other.
The conventional sign of the worldview associated with risk in the organisation is its alliance with the word ‘management’.
Human risk-taking stretches the envelope of possibility in both negative and positive directions: negative where harm may occur to the self and others, positive where the outcome is life-enhancing for one or both. Risk management focuses on constraining the envelope’s negative dimensions to within acceptable parameters. It ignores the human dimensions of risk and their implications for our life and positive flourishing. Waves of redundancies and cuts in industry and academia thus unfold with little, if any, real regard for their effects on human lives.
In On Freedom, Snyder writes that ‘[a] free speaker is taking a risk. Those who take risks need solidarity, in the form of protection of the freedom of speech’ (p.264). This is as true of creative risk as political risk (in the regimes described by Snyder, these are often one and the same). Creative arts departments can be a place of solidarity for creative risk-takers. Creative risk is a form of unfettered speech that reaches beyond conventions, beyond what others can predict, or imagine. It is so for those who follow the artistic work over the cliff into unimagined places, and for those who receive it as a gift to their own creative life potential.
For myself as a lyric poet, Emily Dickinson exemplifies both the ‘cliff’ and ‘gift’ aspects of poetry. Reading her has tutored me in stylistic risk.
That is why, as part of my academic research, I recently began to collect AI-generated versions of Emily Dickinson poems (‘AI Emily’) and to analyse these in relation to the original poems. This project is grounded in a desire to illuminate the nature and the stakes of human creative risk.
AI engines generate outputs from statistically likely combinations of words. In the case of the Dickinson corpus, I prompted the engines using some of her own words to seed new continuations of her poems. The resulting versions are strikingly adept in their rhetorical shapes and lexical choices. They are genuinely difficult, as my colleagues have attested, to distinguish from original Dickinson poems. And yet, they are lexical and structural mirrors which entirely lack any concept or experience of lived creative risk.
As I examine this set of AI poems more closely, I hypothesise that I will find that they are more highly predictable, prettier, more positive, and less challenging than the original poems.
I may be proven wrong. But even if the outcome shows few statistical differences in terms of lexical and semantic diversity and complexity and sentiment, there is this: Dickinson experienced these poems as the gift of her own, extraordinary, creative risk-taking. They belong to her Leib. Her creative risk, and her work, cannot be separated, even if characteristics of the work are replicable by algorithmic mirrors.
To return to the Snyder quote, in the context of generative AI, there is a risk that human creative activities will become mere inputs to the process of extracting human knowledge, insight and experience by and for generative AI. The US Authors Alliance described such an exploitative vision succinctly in its March 2025 submission to the US government’s AI Action Plan, where it proposed that ‘[t]his administration should fund creative work on a large scale, in the service of generating data that can fill in any current gaps surfaced by AI developers.’ As creative artists, we risk allowing our risk-averse institutions, on the one hand, and on the other, the billionaire developers of powerful AI, to devalue creative risk and to decouple creativity from its source in human lived experience, the Leib. Neither, I think, is a risk we can afford to accept.
References
Snyder, T. (2024) On Freedom, The Bodley Head, London.
Authors’ Alliance Comment on US AI Action Plan (March 14, 2025). Available at: https://www.authorsalliance.org/2025/03/14/authors-alliance-comment-on-us-ai-action-plan/ (Accessed: 7 April 2025)
Judith Bishop’s third poetry collection, Circadia (UQP) was published in 2024. Judith’s poems are featured on two albums by composer Jane Stanley, Cerulean Orbits (Delphian Records, 2024) and Composeher (Glasgow School of Art Choir, 2024). Her awards include the 2019 NSW Premier’s Literary Awards Kenneth Slessor Prize for Poetry for Interval (UQP, 2018), the Anne Elder Award for Event (Salt Publishing, 2007), the Marten Bequest Travelling Scholarship for Poetry and the Australian Book Review Peter Porter Poetry Prize (2006, 2011). Her essays on poetry and AI have been published in Island (Australia) and PN Review (UK).
Main image: Judith Bishop, by Petr Kuzmin