How AI can help journalists rebuild a fraying connection with their audience

I would like to explain what I think are the three key challenges facing journalism today – before moving on to how AI can be used to address them. But the challenges I’ll discuss are perhaps not those that first come to mind when we think about the state of journalism today. That’s because they specifically relate to journalism’s connection with the public – without which journalism is not able to play the essential role it does in democracy.

Our research at the Reuters Institute examines changes to the profession of journalism, the practices it uses, and the impact of technology on those practices. But much of our work also examines the public’s relationship with news and the news media – how people access news, what they think and feel about it, what they think it does and does not do for them as citizens, and how they see its role as an institution within society.

Journalists have a shared sense of how they feel about these issues – but it is vital to ask whether the public shares this view.

If we think about the very important and very clear economic, practical, and political challenges facing journalism today, and we think about how we would want to these to be solved - or what we think they could look like in an ideal world - I would argue that all of them require not only the existence of an audience for journalism, but also a healthy relationship between journalism and the public.

In other words, and as me and my colleagues have argued before, journalism can only ever exist within the context of its audience.

With that in mind, what can we say about what is happening to news, journalism and its relationship with the public? I would say there are three key trends right now. These trends are overlapping and interrelated, but together they pose a challenge to that connection between journalism and the public. 

News use is changing

The first trend is changing patterns of news access.

We all know by now that traditional, offline forms of news use – particularly printed newspapers, but also television news – are used by fewer and fewer people. Young people in particular are turning to digital, online news access – often via intermediaries like social media, usually accessed on smartphones and other mobile devices.

Here is some data from our Digital News Report – a study of how people get news across 47 media markets. I’m using Germany as an example, but we see many of the same patterns everywhere. The chart shows how weekly TV news use has declined by about 20 percentage points since 2013, and is now used less widely than online news.

We also see a very sharp decline in the use of printed newspapers and magazines, and at the same time the use of social media for news has steadily increased – overtaking print in 2019.

At this point you may be thinking that this is a superficial change – after all, the news itself is still often (though increasingly less so) coming from the same brands as it always has. 

In one sense, this is true. But it is also true that the way people access news seems to be having a big impact upon how engaged people are with news, how much attention people pay to it, and the benefits it gives society and the impact it can have.

There’s considerable evidence to suggest that, although there are some benefits to digital news access, such as increased diversity and choice, people pay less attention to it, consume it less frequently, and benefit from it less.

At the bottom of this chart you can see that the proportion who do not use any of these sources has grown from 1% to 8%, and is steadily increasing in many of the countries we study. In some, it is already over 10%.

Lower interest in news

This brings us to the second key trend: declining interest in news.

We have seen a stark decline in the proportion of people with high levels of interest in news in the last 10 years. In some European countries the proportion with high levels of interest in news has declined by around 20 points or more.

You can see on this chart examples of this trend in the UK, France, Spain, Germany and Poland. In the UK, the figure has gone from 70% in 2015 to just 38% today. From a clear majority to a clear minority.

These trends matter because personal motivation has long been shown by research to be the primary reason why some people engage with news and some people do not.

This is perhaps one reason why our research has also documented an increase in selective ‘news avoidance’ in recent years – with around 40% saying they sometimes or often actively try to avoid the news.

Lower trust in news

The third trend I’d like to highlight is declining trust in news.

Every year we ask people in our survey whether they “trust most news most of the time”, and when we look at the results, we see a decline in the proportion that agree with this statement.

Again, here are some examples from the UK, France and Germany – each of which have seen declines in trust in news since 2015.

The decline in trust is not as dramatic as the decline in interest – and it’s not as consistent, either. There are some countries within Europe – such as those in the Nordic region and the Netherlands - that have even seen small increases in trust. But in most of the markets covered by our research the trend is downwards.

To be clear, this does not necessarily mean that journalism has actually become less trustworthy. In my opinion, although it is necessarily imperfect, journalism is in many respects getting better all the time. And it is clear that declines in trust are influenced by factors outside of journalism, such as criticism from politicians and other elites, for example, who sometimes simply dislike the coverage of them and their actions.

But the reason I think the data on trust is important is that it encapsulates a range of different concerns and negative attitudes people have towards journalism and the media as an institution.

These perceptions may not be accurate or justified – and many of them are not – but this doesn’t make them or their effects on society any less real. Most newsworthy events happen outside of people’s direct personal experience, so a degree of trust is an essential precondition for people to learn about the world from news.

How AI can help

Changing patterns of news use, declining interest in news, and declining trust are the three defining trends for journalism’s relationship with the public today. But how does generative AI fit into this picture and what is its relationship with these challenges?

In many ways, it is too early to say. The technology can be used in an infinite number of ways, many of which are yet to be conceived of. But even now, many newsrooms across the world are integrating AI into their work – whether it is to help them with routine tasks like translation and transcription, or for more ambitious projects around the creation of content.

I’m not going to address the broader question of whether this is right or wrong. Many people have strong views about this, and the concerns about copyright, jobs and professional autonomy undoubtedly deserve serious attention.

Outside of journalism, there are lots of other issues and concerns – including real fears about how generative AI could be used to generate misinformation.

But if we assume that the adoption of AI by journalism will continue, it’s important to think how we can use it to address the three key challenges I’ve mentioned.

Let’s think about these in reverse order, and start with trust in news.

In my view, this is the hardest issue to address with AI. It is much easier to see the challenge here, and harder to see where the opportunity might be.

One key reason for this is that research suggests that people’s trust in news is an extension of how they feel about politics and society in general. This means that people’s trust in news is shaped by external factors that journalists and the news media do not have any direct control over.

In a separate study of attitudes towards the use of generative AI in news across eight countries, we found that people think that the use of AI in journalism will make news ‘cheaper to make’ but also ‘less trustworthy’. 

The blue bars on this chart indicate the things that people think, on balance, generative AI will improve, and the pink bars indicate things that people think AI will make worse.

Although it may be possible to use AI to adapt content to move the needle towards trust for some people, the main task may be in simply countering the perception that AI makes news less trustworthy, and convincing audiences that it is being used responsibly and for their benefit.

This is why efforts around labelling, explanation about how AI is being used for news, and having a human in the loop, while no silver bullet, are nonetheless important for public understanding, and ultimately, public trust.

What about the declining interest in news?

This is also a difficult challenge and it’s not immediately obvious how AI might fit in. 

As researchers, we probably don’t know enough about what shapes people’s interest in news – but there is evidence that interest is shaped in childhood through socialisation with parents and peers. Therefore it may not be possible to use AI to really reverse a specific individual’s interest in news. But I do think there is the potential for AI to be used to better meet the needs of those who are less interested in news.

For example, what if generative AI could be used to take a news article written by a journalist and create an additional version of it that is better suited to people with low interest or lacking background knowledge of complex stories – a version that does not assume the reader had knowledge of what has caused long running conflicts, or does not assume they are familiar with insider political jargon?

And when it comes to news avoidance, news avoiders themselves tell us that one of the main reasons they avoid news is that they do not feel the news is relevant to their lives. Can AI help journalists craft articles that explain how news might affect them – or even just remind journalists that this matters for some people?

Finally, let’s consider how AI can help journalists address the challenge of changing news access. 

To me it is clear that the changing patterns of news access that I spoke about earlier are not going to be reversed. In fact, they will likely become even more embedded, and it may be that AI itself will become a key part of how people access news. 

It’s not so difficult to imagine a future where people ask AI chatbots to give them the latest news. In fact, our research suggests that a small minority of 5% have already tried this. But our research also suggests that in their present state, AI chatbots are not terribly good at accurately serving up the latest news from specific outlets.

Perhaps this will be true for some time, and as mentioned earlier, there are important legal, practical and financial issues that must be addressed before this even becomes a viable proposition. Hallucinations are an obvious concern, too. But personally, I am more interested in how it might be possible to use AI to create additional versions and formats that are a better fit for how people access information in the digital world.

For example, taking articles written by journalists and using generative AI in the newsroom, or on news publishers’ own websites, to create summaries that convey the key points – or to convert that text to audio (or vice versa) to fit in with how people prefer to access media today.

Can AI be used, for example, to create versions of articles that meet the needs of citizens with hearing or visual impairment? 

Uses such as this interest me and appeal to me primarily because they are not about using AI to do the work of journalists, and because the technology is being used in the newsroom, they leave journalists, rather than technology companies, in control of the output. Nor are these examples aimed at making the news cheaper to make – something that is of little, indirect public benefit.

On the contrary, they are about using the technology to do more of what journalism does best, to do it better, and in ways that might help repair the fraying connection with the public. 

This is a version of a speech the author gave at the Lorenzo Natali Prize award ceremony, held in Brussels on 26 November 2024.