As news organizations navigate the uses and pitfalls of artificial intelligence, new research indicates that news audiences want their journalists to be transparent about how they use it.
Trusting News, an organization that advocates for journalists to demonstrate their credibility and earn trust, published a study in September that measured news consumers’ trust in AI uses for journalism. Nearly 94% of news consumers who participated in the survey said they wanted the use of AI to be disclosed. More than half said they wanted to know how AI was used in the reporting process and details of how the tools were used.
The study involved 10 newsrooms and 6,000 responses from news consumers. It found that readers have different comfort levels regarding how journalists use AI. While it may not be surprising that 82.1% expressed being “uncomfortable” with using AI to write stories without human review, 16.4% expressed being “uncomfortable” even with spelling and grammar checks. Large chunks of readers also expressed discomfort with using AI voices for newscasters and writing headlines or social media posts without human review.
Lynn Walsh, assistant director for Trusting News, organized the survey. She said she’s been on calls with journalists and news executives for years about AI.
“And to me, the question popping up is, ‘Well, should we disclose this?’” she said. “And I guess I was a little surprised to hear journalists saying, ‘Well, I don’t know, do we? Do we really need to? We’ve been using technology and spell check and some of this stuff for years.’”
She said she was made aware of an initial study that posed as many questions as answers regarding reader trust and AI and wanted to study the topic more carefully.
“We wanted to know if can we disclose in a way that will build trust?” she said. “That’s kind of the big question we want to answer.”
Each newsroom also did one-on-one interviews with 10 members of their communities to dig deeper into the topics.
One of the big takeaways, Walsh said, is that news consumers are much more comfortable using AI when people are still involved in checking for accuracy and fairness.
Another takeaway is that there is more comfort from news consumers when journalists use AI as a tool rather than for producing content. For example, using AI to check spelling and grammar or for language translation and transcription services are seen as tools that people are comfortable with. People were less comfortable using AI for data analysis.
Walsh recommends that news organizations host community meetings to hear readers’ concerns. Artificial intelligence is an umbrella term for many uses of technology. Using AI without informing readers can have negative consequences.
“Let’s say people in the community find out about this and are like, ‘Whoa, whoa, whoa,’ and think AI is terrible. ‘We don’t understand how this works.’ Or, ‘we think you’re just cheating. Why should we trust your content?’ … So that’s why we’re recommending a check-in with your community about how they feel about this technology and what they know. It will help you make decisions, or it should, so you can serve them in a way that will be helpful, and they still trust you and want to consume your content.”
Trusting News put together a trust kit for editors and publishers. It includes what to include in their disclosures and how to engage readers. Soon, Trusting News will recruit more newsrooms that regularly use AI to participate in a cohort to see how readers respond to the technology and its explanations.
Bob Miller has spent more than 25 years in local newsrooms, including 12 years as an executive editor with Rust Communications. Bob also produces an independent true crime investigative podcast called The Lawless Files.
Comments
No comments on this item Please log in to comment by clicking here