Public art and creative intervention on AI

Use art — museum-scale installations, public projections, exhibitions, artist-books, performance work, tactical-media interventions — to make the sensory, moral, and political stakes of AI legible to publics that discursive advocacy does not reach. The vehicle is the art world's own institutional apparatus: major museum shows (MoMA, Barbican, HKW), biennials, artist residencies, university and civic-space exhibitions, and the public-programming register (talks, catalogues, guided tours) that accompany them. Anatomy of an AI System, Trevor Paglen's ImageNet Roulette and training-data portraits, Stephanie Dinkins's community-AI works, Refik Anadol's data-critical installations, Poulomi Basu and Trevor Paglen's Anadolu Agency AI-imaging work, artistic-community responses to Adobe Firefly and Stable Diffusion, are the working exemplars.

An actor chooses this strategy because the AI-good movement's default register — technical papers, policy briefs, litigation filings — reaches an audience of policy-adjacent professionals but bounces off the general public whose political weight the movement ultimately depends on, and art is the register that materialises the abstract: a labour-and-water map of GPT-3, a wall of ImageNet's racial-slur classifications, a projection of the compute-water-cobalt supply chain, a performance of algorithmic-decision refusal. The strategy also holds AI-cultural ground the entertainment-industry adaptation of AI (the Netflix futurism, the vendor-friendly TED talk, the tech-triumphalist documentary) will otherwise occupy alone, and works on time-scales — a piece exhibited in 2019 continues to teach viewers in 2026 — that discursive advocacy rarely matches. Art also reaches institutions the coalition otherwise cannot: museum boards, arts-council funders, and cultural-diplomacy channels whose composition and audience differ from the movement's normal coalition table.

It trades measurable political outcome for cultural-substrate work. The strategy's outputs are not laws, contracts, or vendor decisions; they are shifts in what a general audience can imagine about AI, and that shift is unmeasurable on the timescales the movement's other strategies operate at. The strategy also depends structurally on art-world gatekeeping — curators, institutions, funders — whose politics are neither reliably movement-aligned nor consistently anti-corporate, and whose selection choices can foreclose the strongest movement-critical work in favour of tech-adjacent aestheticisation. And the strategy is chronically capital-intensive on the artist's side and prestige-scarce on the coalition's side: the artist who takes it on carries the practical burden and the institution takes the reputational credit.

This strategy differs from strat-popular-culture-shaping-ai-imagination by institutional register — popular-culture works through mass entertainment (film, television, novels, viral video), public-art works through the museum, exhibition, and biennial system — and from strat-counter-narrative-framing by medium of address (sensuous artefact vs. discursive proposition). It is the strategy the AI-good movement is currently most under-provisioned in relative to its adjacent-movement peers (climate art, racial-justice art, HIV/AIDS art — all deeply institutionalised in their movements, and each producing consequential cultural infrastructure that outlived any single campaign).