Building trust in wartime

How do you maintain audience trust in a wartime news environment when so much of the information landscape is false?

The team at Suspilne Ukraine, Ukraine’s national broadcaster, has faced countless challenges since Russia’s full-scale invasion in February 2022.

There’s the fight to stay on air, broadcasting from bunkers as infrastructure is attacked. There’s the fight to keep journalists safe as they cover an increasingly unpredictable frontline, with drone detectors now as essential as flak jackets.

And there’s the fight to keep audience trust, amid floods of disinformation.

In a recent event hosted by BBC Studios, BBC Media Action brought together colleagues from Suspilne Ukraine, BBC Research & Development and our own AI specialists to share how content provenance technology – which ‘stamps’ content so that audiences can see when and where it was taken – is helping to bridge this trust gap.

What's real is denied

“The challenge we're approaching with this implementation is not that you think. It's not that fakes will be believed. It is that the real things will be denied. Once people accept that anything can be generated, authentic evidence can be dismissed by simply saying the word AI. No forgery is required,” said Kyrylo Iesin, senior product manager at Suspilne Ukraine.

It’s a challenge faced by news agencies all over the world.

“We started working on C2PA (content provenance) because news is under attack,” said Laura Ellis, head of tech transfer and partnerships for BBC Research & Development, which has established pilots of the technology for the BBC.

“Trust is diminishing, and we've seen that on many fronts for so many years and for so many reasons. Mis and disinformation is proliferating, and AI is gaining a firmer and firmer hold on the whole of our information landscape. And the way we access news is changing fundamentally. We're gradually being more and more disintermediated, as people access content directly online instead of through trustworthy platforms like the BBC.

“Trust in news worldwide is now 37%, which is hugely low. It just feels quite scary. Online consumption is dominating in most markets, and people are consuming news out on the internet, where we know it is much less safe than it has traditionally been on our own websites. And 50% of consumers now get their news from those third-party platforms.”

Flipping the information script

In this precarious space in the news landscape, BBC R&D decided to flip the script.

“You could try all the time to detect what's false, but what we did was to stand the question on its head and say, rather than trying to detect what's false, let's actually show where our information has come from and how it was made. So, people then can make a choice," said Judy Parnall, head of standards and industry for BBC Research & Development.

The content credential is a marker in a piece of content that shows where it’s come from and how it was made. Produced by specialised cameras, each piece of content then has a formal certification, which comes up as a ‘CR’ stamp that is linked back to a particular organisation. Each step in the production process – from source to agency to publisher - adds a marker for the content ‘signature’ so that you can see the history of that content.

If the content is altered, that link is broken.

Learning together

BBC R&D is piloting the approach on the BBC’s own website and supports coalition efforts that have successfully established content provenance usage in many other broadcasters – including France Television, the Canadian Broadcasting Corporation/Radio Canada and Media Corp Singapore. Suspilne Ukraine was a logical extension.

“Nowhere is the veracity and trustworthiness of news more important than in the circumstances that Ukraine finds itself,” Judy said.

BBC Media Action has worked with Suspilne Ukraine for more than a decade and has funded this project through support from the UK’s Foreign, Commonwealth and Development Office, including the provision of two specialist cameras.

“This project has been about building long-term relationships and learning together. Trust cannot come from a single technology - it starts from credible original journalism, reinforced by strong editorial standards and transparency,” said Muge Ozkaptan, BBC Media Action’s senior product and AI lead.

“We know that AI and other emerging technologies bring real opportunities, but also real risks. Their value depends on how they are designed, how they are governed and used. Our wider approach is based on the idea that AI usage should be responsible, inclusive, and build around the real needs and real-world realities. That means thinking about privacy, safety and transparency from the start, and it means keeping humans in the loop, particularly where decisions or information could have significant consequences,” she said.

BBC Media Action works in nearly 50 countries around the world and has a network of nearly 600 local media partners. Virtually all are working in environments where disinformation is distorting the information landscape; an important part of the BBC charity’s work to support them is training in responsible use of AI and identifying increasingly sophisticated mis- and disinformation.

BBC Media Action aims to expand this work to more partners and contexts, building on this learning, and working with BBC R&D as a technical partner as funding permits.

'We broadcast reality'

In Ukraine, where countering denial and doubt is an essential part of Suspilne’s daily reality, the content provenance project is an important tool in documenting everyday life in wartime in ways their audience can trust.

“Our mission is not only producing content. We broadcast reality. And we capture moments of living, the everyday life of individuals, communities, and our entire country,” said Victoria Murovana, head of content for Suspilne Ukraine.

“We all know that denial and doubt … is the biggest challenge, not even sophisticated debates, because when the truth is under question, that's the issue. And this is the biggest challenge for the whole world and for Ukraine first.

“Our crews document war, document atrocities, document [attacks on] civilian infrastructure. And then comes the generic message - it's false. It's AI. So now we have a tool to provide the content credential.”

Read more about the project in our press release.