Thirteen Authors Face Backlash for 2022 Google Wordcraft AI Writing Study
Thirteen professional fiction writers who participated in a 2022 Google AI writing study are facing organized backlash, including boycott threats and public name-and-shame lists, after the study was surfaced on social media earlier this month. The participants, recruited under non-disclosure agreements to an experimental program called the Wordcraft Writers Workshop, had been given early access to a LaMDA-based writing tool and eight weeks to produce a short story.
The backlash was triggered on August 6, when author Lily Lachance, who describes herself as an anti-generative-AI absolutist, posted a link to the Wordcraft study on her Bluesky profile. Lachance reportedly discovered the Wordcraft website while browsing Codex, an exclusive online forum for professional speculative fiction authors. Her Bluesky post has since been deleted, but the reaction it set off has continued to spread across social platforms, with users circulating lists naming all thirteen participants and calling for boycotts of their published works. Engadget
The underlying tensions are not new. Much of the anger among writers stems from the practice of technology companies training large language models on vast quantities of copyrighted text, often without the permission or knowledge of the authors whose work was used. The Wordcraft study, conducted in spring 2022, placed thirteen writers — including award-winning science-fiction authors and New York Times bestsellers — in a position that, from the perspective of aggrieved peers, looked like collaboration with a company whose training practices many writers view as exploitative. Engadget
The tool itself, called Wordcraft, was described as resembling Microsoft Word with a chatbot interface on the side, built as a writing-centric front end for LaMDA, Google's then-flagship conversational model. Each participant was given eight weeks to write a short story using the LLM. At least one author used no AI-generated text whatsoever during the exercise. The resulting stories were published on Google's Magenta website alongside a white paper from the researchers, which was also posted to arXiv. Engadget; arXiv
The study was conducted during a window when Google was still openly publishing AI research that competitors could, and did, build upon. By 2023, the company had shifted to a more defensive posture after ChatGPT demonstrated that openly shared research could accelerate rivals. The Washington Post reported in May 2023 that Google had begun restricting the publication of AI research that had previously helped jump-start competitors, a strategic recalibration driven by the competitive shock of OpenAI's ChatGPT rollout.
The Wordcraft workshop occupies an unusual position in retrospect. It was an early, visible experiment in putting generative AI tools directly into the hands of creative professionals, published openly with both artifacts and a research paper, at a moment before the public debate over generative AI and creative labor had reached its current intensity. The same kind of study conducted today would almost certainly face immediate scrutiny rather than a four-year delay between publication and backlash.
What is striking about the current reaction is the retroactive framing. The study was public. The stories were published. The white paper was available on arXiv. None of this was concealed. Yet the gap between the study's quiet reception in 2022 and the furious response in 2026 illustrates how rapidly the cultural context around generative AI has shifted among creative professionals. What was once an academic curiosity has become, for many writers, a flashpoint.
The call to boycott all thirteen participants, including the author who used no AI-generated text, reflects the depth of feeling in the writing community, but it also raises a question worth flagging: whether participating in a disclosed, time-limited research study years before the current backlash materialized constitutes an offense warranting professional consequences. The fact that the study predates the widespread availability of consumer generative AI tools, and that Google's own posture toward open AI research has since reversed, adds layers of context that the current outrage does not fully account for.
The broader context here is a creative community still negotiating the boundary between collaboration and complicity in an industry where the technology itself has moved far faster than the norms meant to govern it. The Wordcraft participants worked under NDAs with a company that has since changed its research-sharing practices, using a tool built on a model that has since been superseded. Whether retroactive judgment of that participation is proportionate is a question the writing community will have to resolve for itself. What is clear is that the fault lines exposed by this episode are not going away, and the next study of this kind, if one is conducted, will be scrutinized in real time rather than discovered years after the fact.
