Large cracks in newsrooms’ critical path to the future

The warning sign over the future of newsrooms is the one found on the London Underground: Mind the Gap.

A future newsrooms study jointly undertaken by the Financial Times and the World Association of News Publishers surveyed 86 countries, including New Zealand. It found four interconnected gaps that are preventing newsrooms from turning plans about their future into a new reality. Many are falling short on strategy, audience trust, capabilities, and skills.

There is no shortage of ideas, and the direction of travel seems to be well understood, but translating that into action is proving harder than many imagined. In the introduction to the report, FT Strategies director Lisa MacLeod issued a stark wake-up call: “Our newsrooms are not well prepared for a disrupted future”.

She put it bluntly, saying that while there was talk of “a grand game about audience-first journalism”, actual workflows tell a different story. Almost two-thirds of newsrooms are still based on single-channel legacy structures. These structures are throttling transformation because content is still produced for a primary platform and then only adapted for other purposes that should have equal (or higher) priority.

Throughout the report there are recurring references to the audience-first journalism that, in most cases, is more a noble aim than a reality. There are exceptions, however, and New Zealand provides one of them.

The report highlights a service-oriented journalism initiative by Stuff that encourages readers to contact the newsroom with problems they want Stuff journalists to investigate. Since it was launched last September, the service has received more than 500 submissions from people who are ‘ticked off’ about something. Solving Stuff has investigated everything from fraud to missing suitcases. The first 10 stories generated more than 420,000 pageviews.

The feature is notable for both its direct reader input and its transparency. It explains the steps a Stuff journalist takes to investigate the problem and, if the problem isn’t solved, it says so. The result is a feature which gives readers a sense of direct involvement that is enhanced by a candid approach on the part of the Stuff journalists involved – Natalie Akoorie and Tony Wall.

It is one example of how audience engagement can work but the problems facing newsrooms are wider and more deep-seated than can be solved with a single initiative. The challenges demand a newsroom-wide end-to-end cycle that is similar to the concept-to-market approach used in product development. Editorial departments are not always good at that sort of thing.

For example, even where strategic priorities have been set, a quarter of newsrooms continue to operate as they always did – making editorial decisions largely on instinct. Less that a third operate in ways that are directly aligned to the strategic objectives they set themselves. Not enough attention is being given to strategic planning principles and critical path analysis.

Strategic decision-making is too narrow. The survey found that in 62 per cent of newsrooms only one or two people were involved in determining long-term editorial vision, usually the editor-in-chief or managing editor. The greatest alignment between strategy and action is in those organisations where non-traditional participants (reporters?) are involved in the process.

Major gaps exist in the development of artificial intelligence in the newsroom. It is recognised as having huge potential but only half of those surveyed think they have the tools necessary to achieve such goals. Four years after the launch of ChatGPT, many newsrooms are still struggling to modernise their technology and adapt their operating procedures.

AI adoption is hampered by a range of factors but poor knowledge combined with scepticism are major inhibitors.

Too many newsrooms face a lack of internal technical skills or AI expertise. ‘Internal’ relates not only to the news media organisation as a whole but to editorial departments themselves. Newsrooms have greater confidence in technology when journalists are involved in shaping decisions on it, and AI literacy and usage is highest in newsrooms that have their own embedded AI teams. The reverse is also true: Fifty seven per cent of respondents to the survey stated they had no AI representation in the newsroom and reported the lowest AI literacy and usage. In other words, the people choosing the technology were not the ones who had to produce journalism with it.

Hence, newsrooms tend to be risk averse in their employment of artificial intelligence. Only four per cent use the agentic or autonomous decision-making skills of AI but almost 80 per cent use it for relatively simple tasks such as transcription and translation that draw on more easily comprehended large language models. Fewer than half the newsrooms reported AI-assisted copy-editing and video production and only a handful allowed autonomous decision-making in those operations.

The combination of inadequate knowledge and risk aversion may be behind a lack of clarity in the purposes to which AI should be put in future newsrooms. Although newsroom leaders talk about ways in which AI can take their endeavours into new realms, the reality is that AI is being used principally as a time-saver. The status quo remains but it is being achieved more efficiently. And that usually translates as fewer people in the newsroom.

As I read through the survey my eyes strayed to my bookcase and a volume to which I have referred more often than its title might suggest. The Hidden Dynamics of Path Dependence is a very useful guide to understanding not only how institutions work but also offers insights into why strategies succeed or fail. Institutions are prone to self-reinforcing-sequences (SRS) that explain why some are resilient. Journalism is awash in SRS and I suspect that is why so much of the newsroom employment of AI has been in making the same institutionalised sequences work more efficiently. That may be why plans for the future become more ragged once we move into unfamiliar territory and why training and upskilling are too ad hoc.

Yet newsroom leaders know that radical change has been forced upon the industry since the dawn of the digital age. Much of that change has resulted in the upending of the news media business model. The survey found the top newsroom goals for 2026 are audience engagement, business sustainability, and audience reach. Fundamentally these are traditional revenue levers and – like ongoing newsroom cuts – do not represent radical responses to radical change.

More promising are three other 2026 goals revealed by the survey, but which have significantly lower newsroom priorities – editorial distinction, editorial expansion, and civic impact. Each of these targets could make innovative use of artificial intelligence but it would require a change of mindset in many newsrooms for that to happen.

The nature of media, ownership models, and support structures must all evolve to ensure that journalism continues to meet the needs of society. Those needs have changed and continue to evolve as both the digital environment and technologies such as artificial intelligence advance at ever-increasing speed.

When I asked ChatGPT whether journalists will still exist in 2036, it was reassuring: “Overall, while the landscape of journalism will change, the core need for quality reporting, critical analysis, and ethical standards will ensure that journalists remain relevant in 2036.”

The FT/WAN survey, however, suggests that a far more coherent approach to newsroom development and the role of AI in journalism is needed before we can look to the future with real assurance.

At least we will be able to benchmark progress. The chief executive of the World Association of News Publishers, Stig Ørskov, promises the survey will become “a recurring global benchmark, a resource that newsroom leaders can return to year after year for rigorous, comparable insights into strategy, operating models, and technology readiness”.

The 2026 survey has put a peg in the ground.