• By César López Linares
• August 10, 2026

Like many other newsrooms, Chilean news site BioBioChile(opens in new tab) has gradually incorporated artificial intelligence into its work.

The outlet is aware of the skepticism surrounding AI, which is why it has doubled down on a principle that has helped it earn and keep its audience’s trust since its founding more than 60 years ago: transparency.

BioBioChile’s journalists regularly publish content created with AI assistance while clearly disclosing when and how the technology was used.

Those disclosures are essential, said Denisse Charpentier, BioBioChile’s deputy editor.

"I think it's important for people to know how we do our work so they can trust it," Charpentier told LatAm Journalism Review (LJR). "For example, hiding the fact that we edited a photo or that an article involved this technology would, I think, generate much greater distrust."

Ingrid Bachmann and Sebastián Valenzuela are two of the authors of the research "The Effects of Generative AI in News on Media Credibility and Selectivity: Evidence from a Conjoint Experiment in Chile."

New research by the Pontifical Catholic University of Chile, Washington State University and the University of Barcelona supports that premise. The study found that, regardless of how journalists use AI, audiences want two things: transparency about its use and human oversight of the technology.

Sebastián Valenzuela, a researcher at the Catholic University, said the study was in response to growing concern about AI's impact on newsrooms(opens in new tab) and declining trust in the media across Latin America(opens in new tab).

"What we wondered was whether journalism, in trying to improve by using more efficient tools such as AI, might end up undermining its credibility, which is already under considerable scrutiny," Valenzuela told LJR.

Ingrid Bachmann, also a researcher at the Chilean university, said the findings reinforce the idea that audiences understand AI should be a tool, not a substitute for journalism.

"In general, people don't really care whether a news outlet uses AI or what it uses it for," Bachmann told LJR. "What they prefer are outlets that are transparent about its use — that is, [outlets] that say, 'We use it, and this is how we use it.'"

What trust looks like

The researchers started by measuring how much people value certain attributes in news organizations.

A total of 2,145 participants in Chile were asked to compare hypothetical news outlets with different AI policies and choose the ones they considered more credible and preferred as a source of news.

The results revealed a clear preference for news organizations that commit to full human oversight of generative AI and are fully transparent about how they use it. Participants also rebuffed newsrooms that take the opposite approach.

The use of generative AI for routine tasks — such as transcribing interviews, correcting spelling or translating text — had little effect on participants' trust. Likewise, AI-generated text and visual content had only a limited impact, although a slight majority of participants preferred that the technology not be used to generate realistic images or journalistic content requiring interpretation or analysis.

"We also found that people are more comfortable with AI being used for softer topics," Valenzuela said. "When it comes to harder news — such as politics, the economy or social issues — they don't like AI being used."

Human oversight and transparency

The findings make one thing clear: Audiences do not want to be misled, Valenzuela said. They also suggest that news organizations need to ensure the quality of content produced with AI.

Bachmann said the findings also reinforce the idea that rigorous journalism — and the editorial standards behind it — continues to matter despite the many changes reshaping the news industry.

"When a news organization is transparent about how it uses AI and also guarantees human oversight, it is signaling rigor and good governance," she said. "Ultimately, it's not about generating content for the sake of generating content, but about taking responsibility for what you produce."

BioBioChile is among the Chilean news outlets that have already turned those principles into newsroom policy. Before adopting AI, the organization considered it essential to establish clear rules and boundaries, Charpentier said.

"That's how we developed our AI policy, which we shared with everyone” in the newsroom, she said. "We also trained our journalists on how to use the tool — where the limits are, what they can do and what they can't — before fully embracing it."

BioBioChile offers several AI-assisted products, including BBCL en 5(opens in new tab), a five-minute roundup of the day's top headlines, as well as AI-generated summaries at the beginning of its articles. In both cases, the outlet discloses that the content was generated with AI and reviewed by a reporter or editor.

The organization also publicly published its AI guidelines(opens in new tab) in its journalistic ethics policy, specifying which processes involve automation while guaranteeing human oversight of all published content.

"I think transparency is about letting people know how you work, how that information was put together, how you arrived at that result — and also making sure they know that if you make a mistake, you'll correct it and tell them about it," Charpentier said.

BioBioChile remains Chile's most trusted digital news platform, according to surveys such as the 2026 Digital News Report, in which 65% of respondents said they trusted the outlet(opens in new tab).

Although the study does not prescribe how news organizations should implement such policies, its authors believe it offers a roadmap to the practices audiences value most.

For Charpentier, BioBioChile's experience points to the same conclusion.

"I also think transparency is the key to trust, and in that sense I believe AI should follow the same path," she said. "As long as you're transparent with your audience, your audience will trust you."

This story was translated with assistance from AI and reviewed by Alejandro Martínez-Cabrera