You are viewing your 1 free article this month.
Sign in to make the most of your access to expert book trade coverage.
Wiley has released a set of publicly available Artificial Intelligence (AI) guidelines and FAQs for authors at this week’s London Book Fair (LBF). Responding to the concerns of academic and trade book authors, Wiley’s guidelines provide suggestions on how writers can use AI in their manuscripts and protect intellectual property in the process.
The publisher’s goal is to “shape” how AI develops and is used by authors, rather than “being shaped by AI”. This is reflected in the guidelines and accompanying FAQs, which guide authors on everything from using AI to developing prompts, and comparing different AI models to disclosing their use of these tools.
The guidelines provide “general principles” for all authors, but according to Josh Jarrett, Wiley’s senior vice president of AI Growth, it’s the FAQs that offer more practical, step-by-step advice for writers. “We hope that they create some standard practices and best practices so that we can all learn together,” he told The Bookseller.
The publisher has been “experimenting” with AI and “participating in that ecosystem” over the last 18 months, and has developed expertise in this area, which has helped the organisation develop these guidelines. This has included licensing content for training AI models and Retrieval-Augmented Generation (RAG) – a framework that seeks to improve the quality of AI-generated answers by using authoritative sources of information.
Discussing the author contracts and terms associated with these deals, Jarrett said: “As a general rule, different forms of licensing are typically captured in author contracts, and we honour those contracts. We pay royalties against those contracts, whenever there’s a licensing arrangement.”
Continues...
When asked about the decision not to adopt an opt-in approach for authors whose work is being used to train Large Language Models (LLMs), Jarrett explained that Wiley’s focus is to adapt and evolve at a speed that matches the rapidly changing AI landscape. Relying on pre-existing contracts allows the publisher to achieve this, he explained: “Large technology companies are trying to move quickly as well, and so you have to be able to move at the pace the of the market […] and not sort of have things done to you.”
Earlier in the week, the publisher also announced a new partnership that provides access to earth science research materials to train and enhance the European Space Agency’s Earth Virtual Expert (EVE) – a LLM designed to improve how researchers and the public engage with information in the fields of Earth observation and Earth science. Wiley signed an agreement with Pi School – an organisation based in Rome, Italy – which is leading the development of the EVE project for the ESA’s Φ-lab.
This partnership reflects a shift from the development and use of AI for more general purposes to the creation of “specific solutions to solve specific problems”. Transparency and attribution are two key considerations in the development of this tool, and the hope is that it will act as “a model partnership”, demonstrating how publishers and authors can work with other organisations to create “narrow tools that are intended to really advance the human condition [and] solve hard problems”.