“How might we use modular journalism and AI to assemble new storytelling formats?” This is the question that the EMEA cohort of the JournalismAI Collab is exploring this year. In this article, Shirish Kulkarni, who is helping guide the cohort representing Clwstwr, shares some takeaways from the first phase of the process, in which participants have been exploring user needs that might be addressed with the support of artificial intelligence.

If we accept (as we surely must) that journalism is a service, then we need to start acting like it. That means the needs of our audiences, users, citizens being the North Star which guides all our decisions.

If we worked at Procter & Gamble or Unilever and completely ignored what our customers wanted, then we wouldn’t be working there very long. It would be completely inexplicable if, instead, we made decisions on what things we made based on how “cool” they would look to other soap makers or their chances of winning an obscure detergent industry award.

Yet, that seems to be how many journalism organisations work, and somehow we’re surprised that the industry is facing an existential crisis?

With the EMEA cohort of the JournalismAI Collab, we’ve been trying to address some of those problems, by focusing deeply on user needs and approaching audience research with an open mind. That means really listening to what it says, rather than simply cherry-picking the things which support the work we’ve already decided to do.

We set the group the task of gathering some meaningful user needs data – either from face-to-face or remote interviews, resources they already had in-house or from research available in the public domain. It was a fascinating exercise that threw up insights which many of us knew instinctively but had never seen so clearly expressed, but also things none of us had thought of.

The cohort (a wide cross-section of organisations from across Europe, North Africa and the Middle East) has very different audiences, and works in diverse contexts, but it was interesting that a number of common themes came up.

The insight that struck me most forcibly was that audiences we don’t normally reach WANT to be informed. What that says to me is that people want to know what’s going on in the world, they want to understand and engage in democratic discourse but yet they don’t find anything in traditional journalism which helps them do that. That is a damning indictment and necessitates deep reflection in the industry, not just a reflexive blaming of the audience for not being “sophisticated enough”. Perhaps, just maybe, it’s our journalism that’s the problem.

Many of the teams looked closely at the needs of younger audiences and turned up subtle and important insights. For some time there’s been an accepted wisdom circulating in the journalistic ether that young people want more video, less text and they want a faster experience because they have short attention spans. Those myths were exploded by almost all the actual research our teams surfaced.

Firstly, of course, what they found was that young people aren’t a homogeneous entity who all want or need exactly the same information in exactly the same way. On the other hand, they’re not a completely different “species” to older people. Many information needs are actually fairly common across demographics. Both qualitative and quantitative data showed that young people are interested in very similar things to older readers and viewers – finance, real estate, macroeconomics, technology, and even wine! That clearly suggests that we shouldn’t instinctively “dumb down” our journalism for younger audiences but instead should “talk up” to them.

What our group’s research largely found was that young people don’t have a particular predilection for video, and actually far prefer easily digestible text. What’s important to them is that content is presented in a way that is optimised for mobile platforms. It might be a mixture of text, images and audio, but what’s crucial is that it works well on a phone. In that same vein, they also loved longer podcasts that gave them a more in-depth and comprehensive understanding of a story or issue.

In fact, almost all our groups found clearly expressed needs for greater contextualisation and background to stories. The “explainer” is somewhat looked down upon in the industry, because it doesn’t fit the model of breaking or “moving” news which feels most exciting to journalists. Yet our actual audiences LOVE having stories fully explained to them and keep asking for more. They particularly want to know why they should care about stories. That means understanding the impact it will have on them, their families and their communities – what I call “Orientation”. This is, for me, the key purpose of journalism.

On that theme, users surveyed by one of our groups criticised a lack of “depth” in reporting. As journalists, we often equate “depth” with “length” but I wonder if that’s a convenient but dangerous misunderstanding. Taken as a whole, the research we shared actually suggests that the “depth” users are looking for is actually just a broader, contextualised picture of a story, rather than just the “on the day” development which can often be incomprehensible to readers or viewers who may not be consuming news every hour or even every day.

They do want to be “involved” in the news though, to be able to express their own agency. Some groups found that audiences wanted more control of the content they received and how they consumed it. They wanted to be able to make just the right amount of choices – not to be completely passive, but also not to have to get involved in too many interactives or complicated configurations. Essentially, throwing all our written, visual and interactive tools at a storytelling format, all at the same time, will rarely work for the people you want to actually enjoy it.

That’s evidenced most tellingly in user responses to Snowfall-type scrollytelling formats packed full of text, images, interactives and links. Journalists love making them, because they make us feel good, elicit praise from our bosses (and sometimes win awards). However, stories like these often get very poor ACTUAL engagement, and in fact the “every storytelling tool in the box” approach probably leads to worse interaction than a straightforward article.

This tells us that maybe our journalists’ instincts are not as good as we think they are. Audiences are very clearly telling us what they want, but we’re just not listening.

That’s largely because the things we like to make are often not the things our audiences want to read, watch or listen to. The teams in the JournalismAI Collab are already demonstrating how keen they are to change those habits, particularly in the service of better meeting the needs of the many people and communities that we’ve failed in the past. What we hope to create is new forms of storytelling that use AI to provide better information, at scale, to people who’ve not previously found what they were looking for in traditional journalism.

The choice is a simple one: either we start genuinely listening to users and produce the things they want and need, or we continue doing exactly what we’ve been doing for the last 100 years and preside over the not so gradual decline of our industry. It’s a stark question, but one which has an obvious answer.

Sign up for the JournalismAI newsletter to stay up-to-date with all the news about the 2021 Collab Challenges.
JournalismAI is a project of POLIS, supported by the Google News Initiative.