by Shelley Fairweather-Vega, Literary Division Administrator, 2022-2025
In May this year, the ATA Board approved and published a report compiled by the AI Task Force addressing changes taking place in the translation and interpretation industry thanks to new Large Language Model (LLM) technologies, popularly known as “AI.” The statement was aimed primarily at consumers of T&I services who might be tempted to replace their human experts with LLMs, and it laid out many of the pros and cons of the new technologies and the role that human professionals still ought to play in language work. But, as several Literary Division members noticed, the statement did not mention any concerns specific to translators who work in the publishing world or translate creative texts.
On July 1, 2025, ATA President Geoff Koby confirmed that no literary translators had been members of the AI Task Force. Koby recognized that our concerns about the statement not addressing literary translation were valid, and he invited the Literary Division to submit a revised statement. An invitation went out to Leadership Council members, and by the end of the same day, we had a new version of the statement ready to propose. The ATA Board approved our revisions in full at its August meeting and published the updated version online on August 15. You can read the updated statement.
The Literary Division’s revisions introduced language about two main areas of concern: style and intellectual property. We inserted a sentence about the risks of using AI translation for work that demands stylistic sensitivity: “For creative works, AI-generated translations risk distorting and flattening the style of the original, misrepresenting the author’s artistic approach.” We also added the phrase “Intellectual Property Violations” to the list of “Dangers of AI” that appears as one of the statement’s bolded headings. Finally, to supplement an existing comment that using AI engines for translation might compromise confidential data, we added language about how AI tools may have been trained in violation of copyright laws (lawsuits on the topic are ongoing). Specifically, we wrote, “When LLMs are trained on copyrighted material, often obtained from websites that have in turn obtained those texts illegally, the intellectual property rights of the authors and translators of those texts are ignored and violated. As AI systems grow, more protected material is at risk of being absorbed and put to uses for which it is not licensed or intended.”
Translators’ copyrighted works are among those that have been used to train AI engines without notice, permission, or compensation. That has been a constant subject of discussion in other translation organizations, and we are very pleased that ATA has now also spoken up for the rights of book translators as AI sweeps through the publishing industry.
Literary Division Administrator Shelley Fairweather-Vega says she is also excited that the ATA Board acknowledged its error in failing to solicit opinions from ATA’s literary translators while the Task Force was doing its work. She observed that LitDiv leadership often feels literary translators’ viewpoints are underrepresented, as they are a minority within ATA, and said it was gratifying that ATA was willing to correct that oversight in this case.
Past AI Task Force Chair Analia Bogdan writes that she is impressed with the work the Literary Division does and finds our perspective to be valuable. “Rest assured, we’ll make sure Literary Division members are included in future discussions and work,” says Bogdan. As a Division, we will need to stay engaged with the ATA Board and its committees and working groups to make sure book and literary translators are included in ATA-wide projects, regarding AI as well as other topics. Our success this summer proves that it can be done.
Read other organizations’ statements on AI and literary translation:
-New model contract language and advocacy efforts for literary translation and AI by the Author’s Guild (https://authorsguild.org/advocacy/artificial-intelligence/)
-Statement from CEATL (https://www.ceatl.eu/tools-of-the-trade/statement-on-artificial-intelligence)
-A “slow translation” manifesto from ITI (https://www.iti.org.uk/discover/policy/slow-translation-manifesto.html)
-Society of Authors statement on AI (https://societyofauthors.org/where-we-stand/artificial-intelligence/)
Leave a Reply