Are human journalists truly irreplaceable? How to safeguard public interest journalism in the age of AI

Much of the media industry is imaginatively stuck in regards to AI. As David Caswell suggested in this essay, the industry’s vision of AI tends to privilege “the importance of journalists in the information ecosystem over the wants and needs of citizens, customers and societies”, ultimately keeping us from reimagining our products and processes in ways that benefit the people we exist to serve.  

This essay outlines a future in which journalism continues to reach audiences and matter to people. Its focus is neither on the economics nor on the experience of that relationship, but instead on news production, and specifically on the role of human workers. So what should humans in journalism do?

1. Are humans irreplaceable? 

A near dogmatic belief in irreplaceability prevents media leaders from rigorously identifying which human contributions are actually unique and worth preserving (or perhaps evolving). As AI technologies improve, there will be fewer tasks that depend on human capabilities. But uniquely human capabilities will continue to exist, and the continued legitimacy of journalism as an institution may depend on how well we identify and prioritise them.

Newsrooms often treat “responsible AI” as synonymous to human oversight. Many newsroom AI policies reflect this, emphasising supervision, but they rarely expand on the practicalities of such values. 

The assumption that responsible AI in journalism must invariably require a human-in-the-loop, though, overlooks the possibility of alternative configurations of human-AI workflows. While the model may give the illusion of accountability and control, it presupposes a configuration in which humans act as arbiters of machine output, which is both conceptually and practically complicated. 

Oversight does not scale with system complexity, and it can’t be applied to dynamic, personalised content experiences. At some point, being “in the loop” stops having anything to do with meaningful agency and instead serves as what Madeleine Clare Elish has described as moral crumple zones, with news workers taking the fall for errors beyond their control. 

Perhaps most significantly, framing humans’ primary role in an AI-driven future as bearers of accountability leaves much of human potential largely unfulfilled.

While it is comforting to feel irreplaceable, I believe we have an obligation to imagine new human-AI configurations.

2. How to design a human-centric AI strategy

In my ongoing doctoral research at the Royal Institute of Technology in Stockholm, and most recently as a Visiting Fellow at the Reuters Institute, I’ve studied a phenomenon I frame as hybridisation. Best understood as the sociotechnical entanglement of humans and machines in contemporary journalistic work, hybridisation presupposes a world where humans and AI collectively produce journalism.

In my conceptualisation of hybridisation, I distinguish between first- and second-order AI affordances. First-order affordances are tied directly to the capabilities of technology, enabling gains such as efficiency and scale. Second-order affordances, by contrast, concern distinctly human forms of meaning-making, judgment, and ethical reasoning. 

Building on this distinction, I differentiate between machine-centric and human-centric hybridisation: 

• Machine-centric hybridisation prioritises the realisation of first-order affordances, where AI is primarily used to substitute, automate, or streamline tasks. 
• Human-centric hybridisation realises both first- and second-order affordances, with the implementation of AI reconfiguring professional roles in ways that foreground human interpretative, ethical, and creative capacities.

Imagine a case where a newsroom deploys a network of AI agents to cover corporate earnings: pulling filings in real time, generating articles, distributing them instantly across platforms. The system is optimised for speed, scale, and cost efficiency, with journalists positioned mainly as liability bearers in case anything goes wrong. 

Here, hybridisation is machine-centric: first-order affordances such as automation and scalability dominate, and humans remain largely peripheral to the news production process.

Now imagine the same beat organised differently. Agents still generate baseline reports, but their role in the workflow is distinctly different: flagging unusual patterns, suggesting connections to wider economic or political developments, preparing summaries for journalists. Human reporters then interpret significance, decide where deeper investigation is warranted, and engage in human-to-human interviews and analysis to uncover the nuances and unexpected impacts of the story. 

In this configuration, hybridisation is human-centric: alongside efficiency gains, second-order affordances emerge as moral judgment, narrative construction, and editorial prioritisation are actively foregrounded, and the socio-technical system of news production is organised to support these capacities. The value of the output is not measured in quantity but quality – which in turn may build on dimensions like uniqueness, plurality of perspectives, or depth. 

3. Many newsrooms are not optimised for what humans do best

Many news organisations likely believe they are pursuing an AI strategy that leads toward human-centric hybridisation: a future where technology augments judgment, creativity, and editorial responsibility. 

In practice, however, the broad industry trajectory suggests something quite different. What is being built – incrementally, often unintentionally – is a form of machine-centric hybridisation. Workflows are optimised for what machines do well: speed, scale, pattern recognition, cost efficiency. Humans are then positioned around those systems, adapting their tasks, roles, and decision-making to fit the logics of machines. 

The consequence is a subtle but significant inversion: rather than engaging in uniquely human activities, work is reorganised to fit machine-driven processes. And once that inversion is embedded at the infrastructural level, it becomes increasingly difficult to reverse.

This mismatch between intent and reality is troublesome. Failing to design for human-centric hybridisation in journalism creates serious institutional risks for news organisations, including the erosion of true editorial accountability and the hollowing out of professional judgment. 

4. What is uniquely human journalism?

In principle, anything that can be reduced to data – structured, stored, computed – can be absorbed into a machine-driven editorial process, and that process is improving rapidly. It is not far-fetched to imagine systems that ingest signals from the world, verify them probabilistically, detect any patterns, and generate narratives that meet or exceed current standards of speed, accuracy, and even stylistic coherence. 

Large parts of what we recognise as journalism today (such as information gathering, synthesis, and distribution) are already partially automated. Extending that pipeline further does not require a conceptual leap. It only requires continued iteration.

But journalism has never been just a workflow. It holds a set of judgments: what is worth noticing, what is meaningful, what is fair, what is true and relevant in context and not just in data. These are not computational problems, even if they can be approximated statistically. As Shuwei Fang points out in this essay, what remains a uniquely human value in the future information ecology may be defined by what resists digitisation. 

Capabilities often framed as uniquely human today – such as those captured by the EPOCH framework (Empathy and emotional intelligence; Presence, networking, and connectedness; Opinion, judgment, and ethics; Creativity and imagination; Hope, vision, and leadership) – may increasingly be approached computationally as AI systems become more sophisticated and socially embedded. This does not mean machines become human(like), but it does require us to continuously re-evaluate which activities remain distinctly human, and which can be sufficiently simulated, replicated, or operationalised by technical systems.

5. What’s at stake?

Some might argue that the way for news publishers to differentiate in the age of AI is to return to human-only workflows. But opting out of technological development is not a viable strategy for those aiming to serve broad publics. At best, it positions journalism as a premium, artisanal product accessible to affluent audiences. At worst, it ignores (and possibly, further alienates) the millions of people who don’t have the financial or cultural capital to attain what is often referred to as ‘premium’ journalism.

The opposite of ‘hand-made’ is not an attractive future, either. Machine-centric hybridisation processes risk reducing journalism to a process optimised primarily for scale, growth, and efficiency rather than public value. As AI systems are tasked with larger roles in producing, selecting and distributing news, human journalists may be pushed to the sidelines, less able to shape narratives, challenge assumptions, or exercise contextual and ethical judgment. 

In organisations driven by platform logics and productivity metrics, this can lead to increasingly homogenised content, weakened accountability, and a narrowing of journalism’s civic mission. The danger is not automation itself, but the adoption of an industrial value system where speed and optimisation consistently outweigh human expertise, editorial independence, and public value.

The question is not whether to engage with AI, but how. Intentional, human-centric hybridisation may be the only credible alternative to drifting into machine-centric systems that gradually displace the very capacities journalism depends on.