christian-lue-7dEyTJ7-8os-unsplash.jpg
Credit: Christian Lue on Unsplash

When talking about artificial intelligence in the newsroom, there is too much focus on the technology and not enough on what it actually does. We want to help journalists, technophiles or technophobes, to explore this topic in an accessible way. So we are launching a new series that brings stories from your peers who work with editorial robots.

When covering stories like the HGV driver shortage, regional inequalities, or the latest covid rules, most news organisations focus on the national agenda. This makes sense but journalists sometimes create an impression that crises only happen in their country, failing to put stories into a much-needed perspective.

Since July this year, ten public service broadcasters from across Europe came together to launch a collaborative news service 'A European Perspective' that allows them to republish international stories in the language of their national audience.

The project is led by the European Broadcasting Union and its members from Belgium (RTBF), Finland (YLE), France (France Télévisions), Germany (BR/ARD), Ireland (RTÉ), Italy (RAI), Portugal (RTP), Spain (RTVE) and Switzerland (SWI swissinfo.ch) as well as ARTE, the Franco-German broadcaster, take part.

Fast story-sharing across borders always hinges on a speedy and accurate translation. This is where machine learning comes in - with an automated translation tool, a story published by one broadcaster can be republished by another in a different language in under one hour. This means that Finnish readers can follow the latest developments in Spain and Italians can access stories from Germany almost immediately.

Here is how it works: a broadcaster in, say, Portugal, publishes a story on their website. The Portuguese editor approves the story for sharing and then another member can either request a translation in their language, or use the article straight away if it has already been translated in their language. A human from the original publisher gives the translation a once-over to make sure it is accurate (which makes multilingual journalists incredibly valuable), and the story is ready for republishing by a partner, usually after another quick check.

[Read more: I am a community journalist: what can AI do for me?]

"It’s interesting because you can see similarities in areas like health service or education," says Fiona Hearst, a multimedia journalist at Irish broadcaster RTÉ, who picks the stories from around Europe that appear daily in 'A European Perspective' section on the RTÉ News website. She is sometimes surprised at seeing that countries are battling with the same problems and governments are often in a pickle for the same reasons, which helps broaden the picture.

To make editors' jobs easier, the tool also allows them to filter news stories by topic, which comes in handy when following a specific beat. For instance, when several countries were hit by severe floods last summer, Hearst shared human stories that would not come through wires.

"Personal hits home a lot more, like people living in a caravan on the side of the road because their home was gone. People can relate to these stories."

The AI-powered translator is pretty good, she said, and it also automatically transcribes and translates video and audio content, which is a massive time-saver for the broadcaster. All that is left to do is to double-check the accuracy of the transcription and the material is ready to use.

However, the tool is less efficient when "listening" to people with thick regional accents, Hearst added. That said, even human journalists can struggle here sometimes.

Once the article is translated and published, the audience can also access the original version, which can be valuable for those who are learning a foreign language. Not only do they get access to a topical story written by a native speaker, but they also learn about a foreign country.

[Read more: I am an editor: what can AI do for me?]

The initiative also offers a good exposure for RTÉ as around 10 of its articles are republished by broadcasters in other countries every day.

All publishers retain their autonomy though and no story can be shared without the approval of the original broadcaster. The differences in the international media law sometimes pose challenges, as not all content, especially audio and video, can be republished in another country. But this gets resolved by human editors.

The Swiss perspective

Switzerland's online news and information service SWI swissinfo.ch, which publishes national and international news in 10 languages, is finding other interesting ways to use the tool.

Dale Bechtel, executive editor, format development at SWI swissinfo.ch, said that the main goal for the broadcaster was to keep up with the global perspective so publishing content from other countries was a "no-brainer."

"Although it sounds like a cliche, we live in such an interconnected world we cannot just provide audiences only with information about where our country sits. We need to know what's going on in other countries," he added, citing topics that still dominate the news agenda like covid or climate change, but also national stories like education.

To add this international perspective to local stories, the team experiments with boxouts at the bottom of an article that broaden the debate on a subject like education inequalities during the pandemic with content from different countries.

However, there is another scope to use the technology in Switzerland which has four official languages. Bechtel is exploring the possibility to share not only international content but also stories from other parts of the country to keep local audiences informed about what is going on in other regions.

[Read more: I am a local journalist: what can AI do for me?]

Although we often think that all Swiss are bilingual, that is not always the case and following news content in another language can be challenging. In the absence of accessible stories from other regions, French-speaking Swiss will often tune in to news channels from France, German-speakers may follow news in Germany and so on. Although this is a great advantage for reading international news, readers and viewers may miss out on many important issues unfolding in their own country.

The main challenge of working with the tool is ensuring correct translation, said Bechtel. This is not so much about AI’s capabilities to translate individual words but about their usage. For instance, a ministry can be called by different names in different countries while performing the same function. So a human editor with journalistic and regional know-how is always required to check the accuracy of the translation but also the correct use of terms or concepts.

While editors do verify translation, they rely on mutual trust when it comes to the accuracy of shared stories. Given that all members are committed to public-interest journalism, their fact-checking standards are generally high. However, nothing is perfect and the tool has an in-built system for flagging up any errors, adding corrections, and unpublishing stories if needed. Because trust is so important, any news organisation willing to join the project has to be unanimously approved by all members.

The pilot project runs until January 2022 and then the members need to decide how to continue. After that members have to decide whether to continue.

Both Hearst and Bechtel would like to see the project grow and explore the potential for wider collaboration between public broadcasters in Europe.

This series is supported by United Robots and Utopia Analytics. Neither of them is involved in the editorial process at any stage.

Utopia Analytics is a Finnish company that enables automated moderation of reader comments and cuts down the publishing delay. Inappropriate behaviour, bullying, hate speech, discrimination, sexual harassment and spam are filtered out 24/7 so teams can focus on moderation policy management.

United Robots AB is a Swedish technology company working in automated editorial content. The company leverages structured data to provide publishers with automatically generated content about sports, real estate, traffic, weather, local businesses and the stock market.

Free daily newsletter

If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).