Some creatives and academics are rejecting generative AI at work and at home on environmental and ethical grounds, but realize they may end up using it anyway
www.theguardian.com/technology/ 2... Joel S. / @joelhs : “The algorithms say ‘More of the same’, because it's all they can do."" www.theguardian.com/technology/ 2... Margot Finn / @eicathomefinn : 'The novelist Ewan Morrison was alarmed, though amused, to discover he had written a book called Nine Inches Pleases a Lady....he had asked ChatGPT to give him the names of the 12 novels he had written. “I've only written nine,” he says. “Always eager to please, it decided to invent three.” Michael M. Hughes / @michaelmhughes : I count myself among them. I have zero reason, or inclination, to use LLMs. — www.theguardian.com/technology/ 2... [image] @hypervisible : “I read because I want to understand how somebody sees something, and there's no ‘somebody’ inside the synthetic text-extruding machines.” Mastodon: @emilymbender@dair-community.social : I appreciate this piece, but I want to correct the record on one point. I don't talk about LLMs as making “collages” but rather as making papier-mâché, and the difference matters! — https://www.theguardian.com/ ... [image] See also Mediagazer
Context & Ripple Effects
The resistance described here extends an established debate over whether generative text is useful creative assistance or unreliable imitation. Earlier coverage found a writer judged style-matched LLM output hollow or approximate, while Emily M. Bender’s work on LLM limits supplied a vocabulary for distinguishing generated output from human expression.
It also arrives after reporting that AI-written material was already expanding online and reducing demand for some human writing work: the early spread of AI-authored content made individual refusal a professional, not merely personal, choice.
First-order effects
- Creatives and academics who object on environmental and ethical grounds may exclude generative AI from their work and home routines, even as they acknowledge that practical pressures could override that preference.
- The reported false novel titles reinforce a near-term reason for skeptics to keep human verification and authorship central when accuracy or personal voice matters.
Second-order effects
- Employers, collaborators and clients working with AI-averse contributors may need workflows that permit non-AI production rather than treating model use as a default.
- As AI-generated material becomes more common, the contrast between human-authored work and synthetic output can become a more explicit commissioning and audience-choice criterion.
Third-order effects
- If resistance persists alongside reluctant adoption, creative work could settle into a mixed system: generative tools are widely available, but human provenance and editorial accountability carry differentiated value.
- The tension points toward governance centered on disclosure, consent and the environmental and labor trade-offs of deployment, rather than a simple choice between universal adoption and blanket rejection.
The trend: Generative AI is moving from a novelty tool toward a contested workplace default, making provenance, reliability and the terms of participation part of creative production.