Caption Comics Festival is running a "Small Press Versus AI" panel on Sunday, 9 August in Oxford, bringing together artists and legal experts to confront the damage generative AI is doing to the creative community. The discussion, chaired by event manager Amy Letts, features speakers Pete Trainor, Hannah Berry, Karrie Fransman, and KitsuneArt, and aims to show that creatives are not powerless against technology built on their work without consent.
Letts did not mince words about her own experience. "250 of my artworks, that I know of, have been used to train generative-AI. These were scraped off my own website where I have clear copyright notices. I wasn't asked or compensated. Art is a personal and vulnerable process. I feel absolutely violated," she said. The violation pushed her to stop updating her webcomic and shift entirely to physical comics and zines. "The tech bros say 'adapt or die' and I say artists are very good at adapting! We won't let the arts die," she said.
Ethical and legal challenges
Several AI companies face prosecution for sourcing data through piracy and commercial espionage. Training relies on underpaid gig workers, while the public is tested on systems with little regard for risks such as AI-psychosis. The industry is estimated to be $1.65 trillion in debt, and the environmental cost of data centres piles onto an already strained climate.
AI companies claim the largest art heist in history is legal under fair use, arguing that outputs are transformative. But an estimated 60-80% of outputs are obvious derivatives, and overfitting produces direct copies of training data. In the UK, fair use does not exist; the narrower "fair dealings" doctrine covers research, education, and news reporting, which does not describe commercial AI. The Report on Copyright and Artificial Intelligence (March 2026) suggests that future AI companies will need to be transparent about training data and license copyrighted works.
Wider risks beyond the art world
The threat extends beyond artists. Family photos are scraped into training sets just as easily as professional portfolios, putting personal images and likenesses at risk. Scammers exploit these tools to create deepfakes, and women have become targets of AI-generated, non-consensual sexual images. The loss of control over one's own image is a problem that affects everyone.
Protecting your work from AI scraping
One of the panel's goals is to educate creatives on practical defenses. In the UK, artists can opt out of AI training, but most platforms permit it by default. Platforms like Cara forbid AI entirely, while website owners can edit robots.txt to discourage crawlers-though no method is foolproof. Tools such as Nightshade and Glaze show evidence of disrupting scraping, at the cost of slight image distortion, and newer options like ArtShield, Mist, and Anti-Dreambooth are emerging. For those navigating this landscape, AI for Creatives offers guidance on how to respond to these technologies.
The backlash against 'slop'
Generative AI outputs have been labelled slop-a flood of low-quality, probabilistic content that feels shallow and inauthentic. Students at Boston University recently booed a visiting speaker who praised AI. Unlike intentional art, AI images lean toward the average and, because they now make up a huge portion of online data, the technology is effectively eating itself. Letts put it bluntly: "Your stick figure comic is a million times better than AI slop."
Why this matters for creatives
The panel is a direct response to an industry that tells creators their own skills are not enough. It equips artists with knowledge about their legal rights, protective tools, and the collective power of opting out and speaking up. Caption Comics Festival, which runs 8-9 August, believes that original, human creativity cannot be replaced by a machine. Tickets are available for the weekend, and the AI panel takes place Sunday at 2pm.
Your membership also unlocks:
