Language Models (LLMs) and the public launch of ChatGPT in late 2022, followed by a flurry of other generative AI tools, marked a real shift in public interest and in how publishers have been thinking about AI for the production of news content.
Research on the uptake of AI in newsrooms is rapidly growing, (e.g. Beckett and Yaseen 2023; Simon 2024), but we know considerably less about how news audiences might receive these incursions, especially in a context of low or declining trust in news in many countries. We have included questions on the topic in this year’s survey, at a time when audiences themselves are learning and forming their own opinions about AI in general and few will understand how these technologies might be used in journalism specifically.
In the analysis that follows, we supplement our survey data with illustrative quotes and insights from qualitative research conducted by the market research agency Craft with 45 digital news users in Mexico, the UK, and the US. The research took participants on an individual ‘deliberative journey’, capturing their baseline understanding and comfort levels with AI in initial interviews, and then presenting them with news-related AI case uses to experiment with and reflect on. Just as publishers can use AI for different kinds of tasks, audiences approach distinct uses with varying degrees of comfort, more generally onboard with backend tasks and innovation in the delivery of news than with the generation of new content.
