Month: April 2025

Posted on

Anne Lagercrantz, SVT: “Journalism has to move up the value chain”

Anne Lagercrantz is the Director General of SVT Swedish Television. Alexandra talked to her about how generative AI has created more value for audiences, SVTs network of super users, and what will make journalism unique as opposed to automated content generation. 

Anne, many in the industry have high hopes that AI can do a lot to improve journalism, for example by making it more inclusive and appealing to broader audiences. Looking at SVT, do you see evidence for this?  

I can see some evidence in the creative workflows. We just won an award for our Verify Desk, which uses face recognition and geo positioning for verification.  

Then, of course, we provide automated subtitles and AI-driven content recommendations. In investigative journalism, we use synthetic voices to ensure anonymity.  

I don’t think we reach a broader audience. But it’s really being inclusive and engaging. 

In our interview for the 2024 report, you said AI hadn’t been transformative yet for SVT. What about one year later? 

We’re one step further towards the transformative. For example, when I look at kids’ content, we now use text to video tools that are good enough for real productions. We used AI tools to develop games then we built a whole show around it.  

So, we have transformative use cases but it hasn’t transformed our company yet.  

What would your vision be? 

Our vision is to use AI tools to create more value for the audience and to be more effective.  

However – and I hear this a lot from the industry – we’re increasing individual efficiency and creativity, but we’re not saving any money. Right now, everything is more expensive.  

Opinions are split on AI and creativity. Some say that the tools help people to be more creative, others say they are making users lazy. What are your observations?  

I think people are truly more creative. Take the Antiques Roadshow as an example, an international format that originated at the BBC.  

We’ve run it for 36 years. People present their antiques and have experts estimate their value. The producers used to work with still pictures but with AI support they can animate them.  

But again, it’s not the machine, it’s the human and the machine together.  

You were a newsroom leader for many, many years. What has helped to bring colleagues along and have them work with AI?  

I think we cracked the code. What we’ve done is, we created four small hubs: one for news, one for programmes, one for the back office and one for product. And the head of AI is holding it all together.  

The hubs consist of devoted experts who have designated time for coaching and experimenting with new tools. And then there’s a network of super users, we have 200 alone in the news department.  

It has been such a great experience to have colleagues learn from each other.  

It’s a top-down movement but bottom-up as well. We combine that with training, AI learning days with open demos. Everyone has access and possibility.  

We’ve tried to democratize learning. What has really helped to change attitudes and culture was when we created our own SVTGPT, a safe environment for people to play around in. 

What are the biggest conflicts about the usage of AI in the newsroom? 

The greatest friction is to have enthusiastic teams and co-workers who want to explore AI tools, but then there are no legal or financial frameworks in place.  

It’s like curiosity and enthusiasm meeting GDPR or privacy. And that’s difficult because we want people to explore, but we also want to do it in a safe manner. 

Would you say there’s too much regulation?  

No, I just think the AI is developing at a speed we’re not used to. And we need to find the time to have our legal and security department on board.  

Also, the market is flooded with new tools. And of course, some people want to try them all. But it’s not possible to assess fast that they’re safe enough. That’s when people feel limited. 

No one seems to be eager to talk about ethics any longer because everyone is so busy keeping up and afraid of missing the boat. 

Maybe we are in a good spot because we can experiment with animated kids’ content first. That’s different from experimenting with news where we are a lot more careful.  

Do you get audience reaction when using AI?  

There are some reactions, more curious than sceptical.  

What also helps is that the Swedish media industry has agreed upon AI transparency recommendations, saying that we will tell the audience that is AI when it has a substantial influence on the content. It could be confusing to label every tiny thing.  

Where do you see the future of journalism in the AI age now with reasoning models coming up and everyone thinking, oh, AI can do much of the news work that has been done by humans before? 

I’m certain that journalism has to move up in the value chain to investigation, verification and premium content.  

And we need to be better in providing context and accountability.  

Accountability is so valuable because it will become a rare commodity. If I want to contact Facebook or Instagram, it’s almost impossible. And how do you hold an algorithm accountable?  

But it is quite easy to reach an editor or reporter. We are close to home and accountable. Journalists will need to shift from being content creators and curators to meaning makers.  

We need to become more constructive and foster trust and optimism.  

Being an optimist is not always easy these days. Do you have fears in the face of the new AI world? 

Of course. One is that an overreliance on AI will lead to a decline in critical thinking and originality.  

We’re also super aware that there are a lot of hallucinations. Also, that misinformation could undermine public trust, and that it is difficult to balance innovation with an ethical AI governance.  

Another fear is that we are blinded by all the shiny new things and that we’re not looking at the big picture.  

What do you think is not talked about enough in the context of journalism and AI? 

We need to talk more about soft values: How are we as human beings affected by new technology?  

If we all stare at our own devices instead of looking at things together, we will see loneliness and isolation rise further.  

Someone recently said we used to talk about physical health then about mental health, and now we need to talk about social health, because you don’t ever need to meet anyone, you can just interact with your device. I think that’s super scary.  

And public service has such a meaningful role in sparking conversations, getting people together across generations.  

Another issue we need to talk more about is: if there is so much personalization and everyone has their own version of reality, what will we put in the archives? We need a shared record.

This interview was published by the EBU on 16th April as an appetizer for the EBU News Report “Leading Newsrooms in the Age of Generative AI”. 

Posted on

Kasper Lindskow, JP Politiken Media Group: “Generative AI can Give Journalists Superpowers”

Kasper Lindskow is the Head of AI at the Danish Politiken Media Group, one of the front runners in implementing GenAI based solutions in the industry. He is also co-founder of the Nordic AI in Media Summit, a leading industry conference on AI. Alexandra spoke to him about how to bring people along with new technologies, conflicts in the newsroom, and how to get the right tone of voice in the journalism. 

Kasper, industry insiders regard JP/Politiken as a role model in implementing AI in its newsrooms. Which tools have been the most attractive for your employees so far?  

We rolled out a basic ChatGPT clone in a safe environment to all employees in March 2024 and are in the process of rolling out more advanced tools. The key for us has been to “toolify” AI so that it can be used broadly across the organization, also for the more advanced stuff.  

Now, the front runners are using it in all sorts of different creative ways. But we are seeing the classic cases being used most widely, like proofreading and adaptation to the writing guides of our different news brands, for example suggesting headlines.  

We’ve seen growing use of AI also for searching the news archive and writing text boxes.  

Roughly estimated, what’s the share of people in your organization who feel comfortable using AI tools on a daily basis? 

Well, the front runners are experimenting with them regardless of whether we make tools available. I’d estimate this group to be between 10 and 15 percent of newsroom staff. I’d say we have an equally small group who are not interested in interacting with AI at all.  

And then we have the most interesting group, between 70 and 80 percent or so of journalists who are interested and having tried to work with AI a little bit.  

From our perspective, the most important part of rolling out AI is to build tools that fit that group to ensure a wider adoption. The potential is not in the front runners but in the normal, ordinary journalists. 

This sounds like a huge, expensive effort. How large is your team?  

We are an organization of roughly 3,000 people. Currently we are 11 people working full-time on AI development in the centralized AI unit plus two PhDs. That’s not a lot. But we also work for local AI hubs in different newsrooms, so people there spend time working with us.  

This is costly. It does take time and effort, in particular if you want high quality and you want to ensure everything aligns with the journalism.  

I do see a risk here of companies underinvesting and only doing the efficiency part and not aligning it with the journalism. 

Do you have public-facing tools and products? 

In recommender systems we do, because that’s about personalizing the news flow. That’s public facing and enabled by metadata.  

We’re also activating metadata in ways that are public facing just for example in “read more” lists that are not personalized.  

But in general, we’re not doing anything really public facing with generative AI that does not have humans in the loop yet. 

What are the biggest conflicts around AI in your organization or in the newsroom? 

Most debates are about automated recommender systems. Because sometimes they churn out stuff that colleagues don’t find relevant.  

But our journalists have very different reading profiles from the general public. They read everything and then they criticize when something very old turns up.  

And then, of course, you have people thinking: “What will this do to my job?”  

But all in all, there hasn’t been much criticism. We are getting a lot more requests like: “Can you please build this for me?” 

What do you think the advancement of generative AI will do to the news industry as a whole? 

Let’s talk about risks first. There’s definitely a risk of things being rolled out too fast. This is very new technology. We know some limitations, others we don’t.  

So, it is important to roll it out responsibly at a pace that people can handle and with the proper education along the way.  

If you roll it out too fast there will be mistakes that would both hurt the rollout of AI and the potential you could create with it, impacting the trustworthiness of news.  

Another risk is not taking the need to align these systems with your initial mission seriously enough. 

Some organizations struggle with strategic alignment, could you explain this a bit, please?  

Generative AI has a well-known tendency to gravitate towards the median in its output – meaning that if you have that fast prototype with a small prompt and roll it out then your articles tend to become dull, ordinary and average.  

It’s not necessarily a tool for excellence. It can be but you really need to do it right. You need to align it with the news brand and its particular tone of voice, for example. That requires extensive work, user testing and fine-tuning of the systems underneath.  

If we don’t take the journalistic work seriously, either because we don’t have resources to do it or because we don’t know it or move too fast, it could have a bad impact on what we’re trying to achieve. Those are the risk factors that we can impact ourselves. 

The other risks depend on what happens in the tech industry? 

A big one is when other types of companies begin using AI to do journalism. 

You mean companies that are not bound by journalistic values? 

If you’re not a public service broadcaster but a private media company, for the past 20 years you’ve experienced a structural decline.  

If tech giants begin de-bundling the news product even further by competing with journalists, this could accelerate the structural decline of news media.  

But we should talk about opportunities now. Because if done properly, generative AI in particular has massive potential. It can give journalists superpowers.  

Because it helps to enrich storytelling and to automate the boring tasks? 

We are not there yet. But generative AI is close to having the potential for, once you have done your news work with finding the story, telling that story across different modalities.  

And to me that is strong positive potential for addressing different types of readers and audiences. 

We included a case study on Magna in the first EBU News Report which was published in June 2024. What have your biggest surprises been since then? 

My biggest positive surprise is the level of feedback we are getting from our journalists. They’re really engaging with these tools. It’s extremely exciting for us as an AI unit that we are no longer working from assumptions but we are getting this direct feedback.  

I am positively surprised but also cautious about the extent to which we have been able to adapt these systems to our individual news brands. Our tool Magna is a shared infrastructure framework for everyone.  

But when you ask it to perform a task it gives very different output depending on the brand you request it for. You get, for example, a more tabloid-style response for Ekstra Bladet and a more sophisticated one for our upmarket Politiken.  

A lot of work went into writing very different prompts for the different brands.  

What about the hallucinations everyone is so afraid of? 

This was another surprise. We thought that factuality was going to be the big issue. We had many tests and found out that when we use it correctly and ground it in external facts, we are seeing very few factual errors and hallucinations.  

Usually, they stem from an article in the archive that is outdated because something new happened, not because of any hallucinations inside the model.  

The issue is more getting the feel right in the output, the tone of voice, the angles that are chosen in this publication that we’re working with – everything that has to do with the identity of the news brand.  

This interview was published by the EBU as an appetizer for the News Report “Leading Newsrooms in the Age of Generative AI” on .8th April 2025.

Posted on

Managers, Talk About Your Fears!

Supposedly, the days are over when newsrooms resembled macho hives, war reporters were the cool guys with armoured souls and journalists could only guess a colleague’s burnout by the length of a sick leave. At least that’s how Phil Chetwynd, Global News Director at the AFP news agency, sees it. While 10 to 15 years ago newsrooms were dominated by a culture of “don’t ask, don’t tell”, since then not only has the discussion about mental health in the workplace reached a different level; appropriate structures have also been put in place, Chetwynd said on a panel at the International Journalism Festival in Perugia. This was necessary, he added, as journalists worldwide were facing unprecedented pressure.

In fact, just scrolling through the festival program provides a good overview of all the pain points: Eroding business are squeezing budgets and AI is accelerating the pace of innovation, while authoritarian politicians and their vassals are discrediting journalism, threatening journalists and sowing mistrust or even hostility towards the profession among the population, which is expressed both online and offline. Nevertheless, it is safe to assume that some media executives have not yet understood this with all its consequences.

In fact, business models can only be made resilient if you take care of those who are supposed to do so. And for all the reasons mentioned, the media industry has lost its appeal. Many journalists and media managers are questioning their career choice or leaving for calmer waters if they can. And the next generations don’t even feel tempted to join. In editorial offices that operate according to the old principle “if you can’t stand the heat, get out of the kitchen”, things could soon become breezy. Starting your own business is an alternative. However, it is more suitable for workaholics and self-exploiters, as was made clear on one or two other panels. So even in start-ups, there is no harm in thinking about mental health at an early stage.

Journalist and consultant Hannah Storm – on stage with Chetwynd – has published a book on the subject in 2024. “Mental Health and Wellbeing for Journalists” is based on 45 interviews with media people and trauma experts from all over the world. Key is to create safe spaces for conversations about the topic, said Storm. And anyone who thinks of post-traumatic stress disorder primarily in terms of war correspondents is underestimating the extent of the issue. For example, those who view disturbing footage day in, day out may be more affected than those who work in the middle of the action, said Storm. Vicarious trauma is the technical term for this. Psychologist Sian Williams has developed some recommendations for newsrooms on how to deal with this.

Such guidelines should also be of interest to local editors, as colleagues who frequently report from accident scenes or are confronted with gruesome details when covering the courts generally receive less attention than those who are sent to crisis areas. And fact-checkers are particularly threatened online – to an extent, as Chetwynd reports, that they no longer sign their texts with their names at AFP. The non-profit company The Self Investigation has put together a toolkit on mental health specifically for this group.

For some it’s not the news that causes burnout, but the accelerated pace in newsrooms coupled with economic pressure and concerns that AI could make the job redundant – all this on top of the normal madness of family management that keeps many colleagues busy in key career years. And this is not only true in cultures where self-fulfilment and leisure time play a major role. At a change management workshop for a media company in Malaysia – I led the seminar – participants brought up work-life balance and mental health as top challenges for the industry.  

Emma Thomasson, now a consultant and previously a bureau chief and senior correspondent at Reuters, addressed the topic proactively at the news agency and met with such a positive response that she set up a corresponding internal program. Today, she is involved in the journalist helpline of the Netzwerk Recherche, which offers help to all those who feel that the stress of their job is getting too much for them.

But managers are not doing their job if they leave such offers solely to external bodies and the initiative of those potentially affected. It is important to create a culture in which employees can talk about such experiences without fear, said Chetwynd. This includes managers also showing themselves to be vulnerable. This challenge must be taken up by all those in leadership positions. At AFP, with its 150 offices around the world, that’s a lot of people, but that’s the only way it works. Chetwynd: “The only barriers to a potentially unlimited amount of work are the managers.”

Experts consider two things in particular to be important to alleviate stress: good preparation and regular breaks. Not every journalist, for example, is suited for every assignment simply because of their character and personal history. And managers should address the potential risks proactively before someone takes on a task – be it a disaster assignment or moderating online comments. Adjustment helps. AFP, for example, only lets crisis reporters go to the front lines once they have become familiar with the environment and culture in less exposed roles. And after a few weeks, they must take time off to breathe. This sounds simple, but in many traditional newsrooms as well as in start-ups, it is part of the culture to quietly let workaholics do their thing if they get unloved work out of the way or increase the fame of the brand. This may work in the short term, but in the long run such negligence can be expensive – not to mention the human cost of it. 

Media companies are generally well advised to think proactively about the current and desired corporate or editorial culture, put it in words and communicate it precisely to employees. Lea Korsgaard, editor-in-chief of the widely praised Danish news brand Zetland, presented clear principles on another Perugia panel. Everyone who joins the company gets an hour with her, “and then I explain the culture”, she said, adding: “If you want to create a human-centric product, you need a human-centred culture.” Any news organization who wants to play a part in the battle for talent should listen up.

This column was written for and published by Medieninsider in German (“Chefs, redet über eure Ängste!”) on 15th April 2025.

Posted on

Prof. Pattie Maes, MIT: “We don’t have to simplify everything for everybody”

Prof. Pattie Maes and her team at the MIT Media Lab conduct research on the impact of generative AI on creativity and human decision-making. Their aim is to advice AI companies on designing systems that enhance critical thinking and creativity rather than encourage cognitive offloading. The interview was led for the upcoming EBU News Report “Leading Newsrooms in the Age of Generative AI”.  

It is often said that AI can enhance people’s creativity. Research you led seems to suggest the opposite. Can you tell us about it?  

You’re referring to a study where we asked college students to write an essay and had them solve a programming problem.  

We had three different conditions: One group could use ChatGPT. Another group could only use search without the AI results at the top. And the third group did not have any tool.  

What we noticed was that the group that used ChatGPT wrote good essays, but they expressed less diversity of thought, were more similar to one another and less original. 

Because people put less effort into the task at hand? 

We have seen that in other experiments as well: people are inherently lazy. When they use AI, they don’t think as much for themselves. And as a result, you get less creative outcomes.  

It could be a problem if, say, programmers at a company all use the same co-pilot to help them with coding, they won’t come up with new ways of doing things.  

As AI data increasingly feeds new AI models, you will get more and more convergence and less improvement and innovation.  

Journalism thrives on originality. What would be your advice to media managers? 

Raising awareness can help. But it would be more useful if we built these systems differently.  

We have been building a system that helps people with writing, for example. But instead of doing the writing for you, it engages you, like a good colleague or editor, by critiquing your writing, and occasionally suggesting that you approach something from a different angle or strengthen a claim.  

It’s important that AI design engages people in contributing to a solution rather that automating things for them.  

Sounds like great advice for building content management systems. 

Today’s off-the-shelf systems use an interface that encourages people to say: “write me an essay on Y, make sure it’s this long and includes these points of view.”  

These systems are designed to provide a complete result. We have grammar and spelling correctors in our editing systems, but we could have AI built into editing software that says, “over here your evidence or argument is weak.”  

It could encourage the person to use their own brain and be creative. I believe we can design systems that let us benefit from human and artificial intelligence.  

But isn’t the genie already out of the bottle? If I encouraged students who use ChatGPT to use a version that challenges them, they’d probably say: “yeah, next time when I don’t have all these deadlines”.   

We should design AI systems that are optimised for different goals and contexts, like an AI that is designed like a great editor, or an AI that acts like a great teacher.  

A teacher doesn’t give you the answers to all the problems, because the whole point is not the output the person produces, it is that they have learned something in the process.  

But certainly, if you have access to one AI that makes you work harder and another AI that just does the work for you, it is tempting to use that second one. 

Agentic AI is a huge topic. You did research on AI and agents as early as 1995. How has your view on this evolved since? 

Back when I developed software agents that help you with tasks, we didn’t have anything like today’s large language models. They were built by hand for a specific application domain and were able to do some minimal learning from the user.  

Today’s systems are supposedly AGI (artificial general intelligence) or close to it and are billed as systems that can do everything and anything for us.  

But what we are discovering in our studies is that they do not behave the way people behave. They don’t make the same choices, don’t have that deeper knowledge of the context, that self-awareness and self-critical reflection on their actions that people have.  

A huge problem with agentic systems will be that we think they are intelligent and behave like us, but that they don’t. And it’s not just because they hallucinate. 

But we want to believe they behave like humans? 

Let me give you an example. When I hired a new administrative assistant, I didn’t immediately give him full autonomy to do things on my behalf.  

I formed a mental model of him based on the original interview and his résumé. I saw “oh, he has done a lot of stuff with finance, but he doesn’t have much experience with travel planning.” So when some travel had to be booked, I would tell him, “Let me know the available choices so that I can tell you what I value and help you make a choice.”  

Over time my mental model of the assistant develops, and his model about my needs and preferences. We basically learn about each other. It is a much more interactive type of experience than with AI agents.  

These agents are not built to check and say, “I’m not so confident making this decision. So, let me get some input from my user.” It’s a little bit naïve that AI agents are being portrayed as “they are ready to be deployed, and they will be wonderful and will be able to do anything.”  

It might be possible to build agents that have the right level of self-awareness, reflection and judgment, but I have not heard many developers openly think about those issues. And it will require a lot of research to get it right.  

Is there anything else your research reveals about the difficulties with just letting AI do things for us? 

We have done studies on decision making with AI. What you expect is that humans make better decisions if they are supported by an AI that is trained on a lot of data in a particular domain.  

But studies showed that was not what happened. In our study, we let people decide whether some newspaper headline was fake news or real news. What we found was when it’s literally just a click of a button to get the AI’s opinion, many people just use the AI’s output.  

There’s less deep engagement and thinking about the problem because it’s so convenient. Other researchers got similar results with experiments on doctors evaluating medical diagnoses supported by AI, for example. 

You are telling us that expectations in AI-support are overblown? 

I am an AI optimist. I do think it is possible to integrate AI into our lives in a way that it has positive effects. But we need to reflect more about the right ways to integrate it.  

In the case of the newspaper headlines we did a study that showed that if AI first engages you in thinking about a headline and asks you a question about it, it improves people’s accuracy, and they don’t accept the AI advice blindly.  

The interface can help with encouraging people to be a little bit more mindful and critical.  

This sounds like it would just need a little technical fix.  

It is also about how AI is portrayed. We talk about these systems as artificial forms of intelligence. We constantly are told that we’re so close to AGI. These systems don’t just converse in a human-like ways, but with an abundance of confidence.  

All of these factors trick us into perceiving them as more intelligent, more capable and more human than they really are. But they are more what Emily Bender, a professor at the University of Washington, called “stochastic parrots”.  

LLMs (large language models) are like a parrot that has just heard a lot of natural language by hearing people speak and can predict and imitate it pretty well. But that parrot doesn’t understand what it’s talking about.  

Presenting these systems as parrots rather than smart assistants would already help by reminding people to constantly think “Oh, I have to be mindful. These systems hallucinate. They don’t really understand. They don’t know everything.”  

We work with some AI companies on some of these issues. For example, we are doing a study with OpenAI on companion bots and how many people risk becoming overly attached to chat bots.  

These companies are in a race to get to AGI first, by raising the most money and building the biggest models. But I think awareness is growing that if we want AI to ultimately be successful, we have to think carefully about the way we integrate it in people’s lives.  

In the media industry there’s a lot of hope that AI could help journalism to become more inclusive and reach broader audiences. Do you see a chance for this to happen? 

These hopes are well-founded. We built an AI-based system for kids and older adults who may have trouble processing language that the average adult can process.  

The system works like an intra-language translator – it takes a video and translates it into simpler language while still preserving the meaning.  

There are wonderful opportunities to customize content to the abilities and needs of the particular user. But at the same time, we need to keep in mind that the more we personalize things, the more everybody would be in their own bubble, especially if we also bias the reporting to their particular values or interests.  

It’s important that we still have some shared media, shared news and a shared language, rather than creating this audience of one where people can no longer converse with others about things in the world that we should be talking about. 

This connects to your earlier argument: customisation could make our brains lazy.  

It is possible to build AI systems that have the opposite effect and challenge the user a little bit. This would be like being a parent who unconsciously adjusts their language for the current ability of their child and gradually introduces more complex language and ideas over time.  

We don’t have to simplify everything for everybody. We need to think about what AI will do to people and their social and emotional health and what artificial intelligence will do to natural human intelligence, and ultimately to our society.  

And we should have talks about this with everybody. Right now, our AI future is decided by AI engineers and entrepreneurs, which in the long run will prove to be a mistake. 

The interview was first published by the EBU on 1st April 2025.

Posted on

Peter Archer, BBC: “What AI doesn’t change is who we are and what we are here to do”

The BBC’s Director of Generative AI talks about the approach of his organization to developing AI tools, experiences with their usage and the rampant inaccuracies AI assistants produce – and what is needed to remedy them. This interview was conducted for the EBU News Report “Leading Newsrooms in the Age of Generative AI” that will be published by the European Broadcasting Union.

BBC research recently revealed disturbing inaccuracies when AI agents provided news content and drew on BBC material. About every second piece had issues. Did you expect this?  

We expected to see a degree of inaccuracy, but perhaps not as high as we found. We were also interested in the range of different errors where AI assistants struggle including factual errors, but also lack of context, and the conflation of opinion and fact.

It was also interesting that none of the four assistants that we looked at – ChatGPT, Copilot, Gemini, and Perplexity – were much better or worse than any of the others, which suggests that there is an issue with the underlying technology.  

Has this outcome changed your view on AI as a tool for journalism?  

With respect to our own use of AI, it demonstrates the need to be aware of the limitations of AI tools.

We’re being conservative about the use of generative AI tools in the newsroom and our internal guidance is that generative AI should not be used directly for creating content for news, current affairs or factual content.

But we have identified specific use cases like summaries and reformatting that we think can bring real value.

We are not currently allowing third parties to scrape our content to be included in AI applications. We allowed ChatGPT and the other AI assistants to access our site solely for the purpose of this research. But, as our findings show, making content available can lead to distortion of that content.  

You emphasised working with the AI platforms was critical to tackle this challenge. Will you implement internal consequences, too? 

Generative AI poses a new challenge – because AI is being used by third parties to create content, like summaries of the news.

I think this new intersection of technology and content will require close working between publishers and technology companies to both help ensure the accuracy of content but also to make the most of the immense potential of generative AI technology.  

So, you think the industry should have more self-confidence? 

Publishers, and the creative and media industries more broadly, are critical to ensuring generative AI is used responsibly. The two sectors – AI and creative industries – can work together positively, combining editorial expertise and understanding of the audience with the technology itself.

More broadly, the media industry should develop an industry position – what it thinks on key issues. The EBU can be a really helpful part of that. In the UK, regulators like Ofcom are interested in the AI space.

We need a constructive conversation on how we collectively make sure that our information ecosystem is robust and trusted. The media sector is central to that.

On the research, we will repeat the study, hopefully including other newsrooms. Because I’m fascinated to see two things: Do the assistants’ performances change over time? And do newsrooms of smaller languages see the same issues or maybe more? 

Do you think the media industry in general is behaving responsibly towards AI? Or what do you observe when you look outside of your BBC world?  

On the whole yes, and it’s great to see different perspective as well as areas of common interest. For example, I think everybody is now looking at experiences like chat assistants.

There’s so much to do it would be fantastic to identify common priorities across the EBU group, because working on AI can be hard and costly and where we can collaborate we should.

That said, we have seen some pretty high-profile mistakes in the industry – certainly in the first 12 to 18 months after ChatGPT launched – and excitement occasionally outpaced responsible use.

It’s also very helpful to see other organizations testing some of the boundaries because it helps us and other public service media organizations calibrate where we are and what we should be doing.  

There are huge hopes in the industry to use generative AI to make journalism more inclusive, transcend format boundaries to attract different audiences. Are these hopes justified?  

I’m pretty bullish. The critical thing is that we stay totally aligned to our mission, our standards, and our values. AI changes a lot, but what it doesn’t change is who we are and what we’re here to do.

One of the pilots that we’re looking at how to scale is taking audio content, in this example, a football broadcast, and using AI to transcribe and create a summary and then a live text page.

Live text updates and pages on football games are incredibly popular with our audiences, but currently there’s only so many games we can create a live page for. The ability to use AI to scale that so we can provide a live text page for every football game we cover on radio would be amazing.

One of the other things that we’re doing is going to the next level with our own BBC large language model that reflects the BBC style and standards. This approach to constitutional AI is really exciting. It’s being led out of the BBC’s R&D team – we’re incredibly lucky to have them.  

Do you have anything fully implemented yet?  

The approach that we’ve taken with generative AI is to do it in stages. In a number of areas, like the football example, we are starting small with working, tactical solutions that we can increase the use of while we work on productionised versions in parallel.

Another example is using AI to create subtitles on BBC Sounds. Again, here we’ve got an interim solution that we will use to provide more subtitles to programmes while in parallel we create a productionised version that is that is much more robust and easier to scale across all audio.

A key consideration is creating capabilities that can work across multiple use cases not just one, and that takes time.  

What is your position towards labelling?  

We have a very clear position: We will label the use of AI where there is any risk that the audience might be materially misled.

This means any AI output that could be mistaken for real is clearly labelled. This is particularly important in news where we will also be transparent about where AI has a material or significant impact on the content or in its production – for example if an article is translated using AI.

We’re being conservative because the trust of our audience is critical.  

What’s the internal mood towards AI? The BBC is a huge organization, and you are probably working in an AI bubble. But do you have any feel for how people are coming on board?  

One of the key parts of my role is speaking to teams and divisions and explaining what AI is and isn’t and the BBC’s approach.

Over the last 12 months, we’ve seen a significant increase in uptake of AI tools like Microsoft Copilot and many staff are positive about how AI can help them in their day-to-day work.

There are of course lots of questions and concerns, particularly as things move quickly in AI.

A key thing is encouraging staff to play with the tools we have so they can understand the opportunities and limitations. Things like Microsoft Copilot are now available across the business, also Adobe Firefly, GitHub Copilot, very shortly ChatGPT.

But it’s important we get the balance right and listen carefully to those who have concerns about the use of AI.

We are proceeding very carefully because at the heart of the BBC is creativity and human-led journalism with very high standards of editorial. We are not going to put that at risk.  

What’s not talked about enough in the context of generative AI and journalism? 

We shouldn’t underestimate the extent to which the world is changing around us. AI assistants, AI overviews are here to stay.

That is a fundamental shift in our information landscape. In two or three years’ time, many may be getting their news directly from Google or Perplexity.

As our research showed, there are real reasons for concern. And there is this broader point around disinformation. We’ve all seen the Pope in a puffer jacket, right? And we’ve all seen AI images of floods in Europe and conflict in Gaza.

But we’re also starting to see the use of AI at a very local level that doesn’t get much exposure but could nevertheless ruin lives.

As journalists, we need to be attuned to the potential misinformation on our doorstep that is hard to spot.  

This interview was published by the EBU on 26th March 2025.