Google's AI Study Scandal: Writers Face Backlash Over Copyrighted Material

Four years ago, Google approached thirteen professional writers to participate in its Wordcraft Writers Workshop. The project involved using an experimental writing tool called Wordcraft, which was built on the LaMDA language model. This model later became the basis for Gemini.

The writers were asked to produce short stories using the interface, and they did so without knowing that it was trained on copyrighted material. Now, some of them are facing backlash from their peers after a social media post highlighted their participation in the study.

Eugenia Triantafyllou, an award-winning author, admitted she didn't know about the model's training data and described Wordcraft as a story editor. She said she would refuse to participate in similar studies today.

Vajra Chandrasekera, another science-fiction author, argued that participants shouldn't be faulted for their past actions, but rather should be judged on their current stance towards AI. The controversy has reignited the debate about whether using AI-generated text is acceptable and raises questions about accountability and transparency in research.