A new Reuters Institute report provides insight into how readers around the world think about the place of generative AI in the news. The survey broadly surveyed the public in six countries: Argentina, Denmark, France, Japan, the UK and the US.
There has been much discussion about the actual impact that generative AI is already having on news publishing, but there seems to be a relatively strong consensus among the general public that generative AI is now widely used and will continue to transform the industry.
On average across the six countries, majority respondents said that journalists currently use generative AI “always” or “sometimes” “with human supervision.” This applies to 11 different editorial tasks, including writing the text for an article, creating AI-generated images to use in place of photos, data analysis, copy editing and headline writing.
Looking ahead, 66% of respondents said they expect generative AI to have a “very large or somewhat large impact” on the news media industry.
But the survey paints a bleaker picture when it comes to perceptions of whether generative AI will make news media better or worse. Scientific research, healthcare, and shopping all received mostly positive responses, while news and journalism received the second-most negative outlook of 14 categories, sandwiched between “equality” and “job stability.”
Despite the overall pessimism, comfort with AI varies greatly depending on what tasks journalists are performing. The survey allowed us to break down the editorial process into smaller pieces and look more closely. On average, people were most comfortable with AI tools being used for basic copy editing (+38%) and translating articles into other languages (+35%). Still, more people were comfortable with AI being used to write headlines (+16) and create audio or video versions of written articles (+15) than not.
But when it comes to producing the core elements of an article, sentiments start to change: -1% were comfortable with AI tools being used to write the body text; -13% were uncomfortable with using AI-generated images when no photo was available; -24% were uncomfortable with creating an artificial author or persona to write and publish the news.
This comfort level also varied greatly depending on the subject matter: people were relatively comfortable with generative AI being used to create stories about fashion, sports, and arts and culture. But the situation changed dramatically when it came to using generative AI to create stories about topics like global affairs, especially politics.
Whether or not respondents were satisfied with articles written using AI tools, the overwhelming majority agreed that AI-generated news is not worth paying dollars, yen, or euros for: Only 8% thought AI-generated news was “worth paying for” compared to human-created news.
Many respondents also said they believe using AI to produce news with “some human oversight” would make it cheaper to produce. 33% of respondents across the six countries said so. This gives us a glimpse into the intentions behind AI adoption in news media. From the outside, readers seem to think that cost savings are the driving force behind introducing generative AI into newsrooms.
News publishers that embrace AI risk eroding the already fragile trust they have built with their readers: Respondents believed that news produced using AI would be less trustworthy overall.
Across 12 different industries, news media was, on average, one of the least trusted to use generative AI responsibly. Only 12% of people in the UK and 30% in Argentina said they “strongly” or “somewhat” trust the news media in this regard. In most cases, news media ranked third behind “national governments,” “social media companies,” and “politicians and political parties.”
