In museums, galleries, and arts organizations, the real debate over generative AI is no longer whether the tools exist. It’s whether institutions want to be seen using them in place of human labor. That question has become sharper in the last year as AI-generated web content has grown more visible: Pew Research found signs of AI authorship in more than one-third of English-language pages published after ChatGPT’s release in a July 2026 sample, while UNESCO and ICOM have been surveying museums worldwide about how AI is being used across the sector. (pewresearch.org)
For cultural institutions, that context matters because writing is not a minor administrative task. Wall text, exhibition copy, fundraising appeals, educational materials, and press releases shape how the public understands art and who gets credited for making and interpreting it. When those words sound machine-made, they don’t just lose texture; they risk flattening the institution’s voice into the same generic prose now flooding the broader internet. As Valentina Di Liscia writes in Hyperallergic, “you’ll know it when you see it: that uncanny cadence that’s overconfident in style yet hollow in substance.” (hyperallergic.com)
That line captures a deeper concern than aesthetics. AI text can be efficient, but efficiency is not the same as authorship. In a field that already asks artists to market themselves relentlessly, the temptation to use generative tools is understandable. Many artists are juggling studio practice with applications, proposals, and public-facing copy, often without adequate support. But the burden lands differently on institutions with staff, budgets, and reputational stakes. When a museum with resources outsources language to a model, it is not solving a labor problem so much as passing it along invisibly. (hyperallergic.com)
Di Liscia argues that the line is clearest where responsibility is highest. “Institutions, on the other hand, especially well-resourced museums, foundations, publications, and other cultural organizations, have no excuse for relying on artificial intelligence for textual materials,” she writes in Hyperallergic. That argument resonates because museums are not merely content factories; they are public-facing authorities. Their language helps define artistic value, historical context, and curatorial intent. If that language becomes standardized, vague, or subtly inaccurate, the institution’s credibility erodes along with its voice. (hyperallergic.com)
There is also a practical issue lurking beneath the ethics: accountability. Human writers can revise, explain, and stand behind their choices; AI systems cannot. In an era when arts organizations are already navigating disputes over authorship, copyright, and creative credit, handing core communications to generative models can create confusion about who shaped the message and why. Pew’s finding that AI content now makes up a substantial share of newly published pages underscores how quickly “synthetic” writing can become normalized online. The art world may be especially vulnerable because its prose often leans on polished abstraction, the very register AI imitates most convincingly. (pewresearch.org)
That’s why Di Liscia’s insistence on original language lands as more than a stylistic preference. “We think someone who writes for a living needs to constantly be thinking about the best way to express complex ideas in their own words,” she notes, quoting Wired’s AI guidelines in Hyperallergic. The point is not that tools have no place anywhere in the process. Research, transcription, and drafting support can be useful. But the final act of writing — the selection of tone, emphasis, rhythm, and meaning — is where institutional voice becomes public trust. When arts organizations delegate that work too casually, they risk sounding less like stewards of culture and more like replicas of the same content machine they claim to resist. (hyperallergic.com)
Originally reported by via hyperallergic.com on 2026-10-20 13:40:08.
Read the full original article here: hyperallergic.com