For many newsrooms today, the arrival of generative AI feels less like a strategic evolution and more like a relentless game of Whack-a-Mole. Whenever a new tool emerges or a competitor announces an AI-driven feature, organizations react in fragments—patching holes and adding plugins without a cohesive map of where they are actually heading.
This reactive posture is more than just an organizational quirk; This proves a systemic failure of strategy. An analysis of 725 cases of AI adoption across media outlets in 80 countries reveals a stark trend: the vast majority of these efforts are designed to automate single, repetitive actions or augment existing human work, but they lack a broader strategic intent. Media companies are using 21st-century technology to simply do 20th-century tasks faster.
Throughout my career reporting from diplomatic hubs and conflict zones in more than 30 countries, I have observed that the institutions that survive profound disruption are those that can distinguish between their operational habits and their core value. For the media, the current AI survival plan for media must move beyond tool-adoption and toward a fundamental reorganization of what it means to be a journalist.
The solution can be visualized through a simple metaphor: the smile. A smile requires two distinct arcs working in harmony. In a newsroom, these arcs represent two non-negotiable priorities: the ruthless reinvention of the human value proposition and the aggressive automation of everything else.
The Adoption Gap: Why Tools Aren’t Strategy
The gap between having access to AI and achieving a transformational business model is wider than most executives realize. When measuring adoption complexity on a scale of 1 to 5, roughly 71% of AI projects in the media are currently stalled at stages one and two—mere access and basic adoption.
This stagnation means that most experiments are failing to produce major efficiency gains or unlock the transformational opportunities of the technology. To move forward, leaders must understand the hierarchy of adoption.
| Stage | Level | Definition |
|---|---|---|
| 1 | Access | Tool exists and team has access, but it is not in regular production. |
| 2 | Adoption | The tool is being used regularly by a specific team or department. |
| 3 | Proficiency | Outcomes are iterated upon and measured via specific KPIs. |
| 4 | Ways of Working | AI is rebuilt into editorial workflows; roles are fundamentally changed. |
| 5 | Reorganising | New business models or entirely new products are enabled by the tech. |
Doubling Down on the Human Moat
If the goal is survival, media outlets must identify the “human moat”—the elements of journalism that no amount of compute power can replace. Whereas some industry leaders believe their specific outlets are safe, the broader reality is that only those who return to the essence of the craft will thrive.
AI investor Sarah Guo suggests that journalists will remain essential due to the fact that the industry requires editorial opinion and human relationships—elements that are not easily proxied by large language models. Similarly, Daniel Hulme, Chief AI Officer at WPP, emphasizes the importance of asking great questions and creating authentic content rooted in empathy and a grasp of the “big picture.”
The consensus among forward-thinking strategists is that media must pivot away from service journalism, evergreen content, and general news—all of which are highly susceptible to automation—and instead invest in original investigations, on-the-ground reporting, and deep contextual analysis.
Essentially, AI cannot witness; it can only recap. There are three domains where human intelligence remains supreme:
- Interpersonal Relationships: Field reporting, building trust with sources, and reading the subtle emotional cues of a protagonist.
- Societal Context: Recognizing patterns in niche communities or historical peculiarities that are often dismissed as “noise” by AI models.
- Authentic Storytelling: Crafting narratives that are grounded and exceptional, providing a necessary antidote to the tide of machine-generated “slop.”
Automating the Mundane to Save the Mission
While the human element is the value, the operational element must be the machine. Stanford professor Erik Brynjolfsson warns against “paving the cow paths”—the mistake of simply layering new technology over ancient, inefficient ways of working. To realize true productivity, companies must break jobs down into individual tasks and outsource every task that does not contribute to the human-led value proposition.
Some organizations are already seeing the results of this approach. The UK-based Newsquest has reported significant productivity gains by training reporters in AI and utilizing proprietary tools, moving from an average of four stories to 30 per reporter per day.
However, the focus must expand beyond mere production volume. Currently, about 66% of AI adoption cases in the media focus on editorial production and workflow, leaving audience experience and discovery largely unexplored.
To bridge this gap, newsrooms should consider forming modest, cross-functional teams comprising engineers, editorial product managers, and audience experts. These teams can conduct rigorous cost-benefit analyses of automation before making irreversible structural decisions.
Learning as Basic Hygiene
The transition to an AI-integrated newsroom is as much a cultural challenge as a technical one. There is a common complaint among managers that journalists are too overworked to learn new tools. However, in an era of recursive self-improvement—where AI models are constantly evolving—learning can no longer be a luxury; it must be treated as basic hygiene.
Some outlets are leading the way in formalizing this education. The Guardian recently rolled out a mandatory AI course for its entire staff, moving beyond a simple list of “dos and don’ts” to explain the underlying science of how AI actually works.
This educational shift is necessary because the boundary between journalism and technical roles is blurring. Just as coding roles are evolving to include more project management and supervision, journalists must evolve into supervisors of machine-produced content, using their editorial judgment—honed by years of mistakes and rewriting—to ensure quality.
there is a need to rediscover “non-air-conditioned journalism.” The last decade of social-media-driven reporting rewarded brief formats and online connection, often at the expense of deep, direct observation. To compete with AI, journalists must return to the grit of the field, cultivating complex relationships and leaning into empathy to tell stories that a machine simply cannot feel.
The danger facing the industry is not a lack of change, but an incomplete one. An organization that automates aggressively but fails to cultivate its human craft is not smiling; it is grimacing. True survival requires both tracks—the technical and the visceral—to run in parallel.
As the industry looks toward the next wave of recursive AI improvements, the focus will likely shift toward deeper integration of AI into audience discovery and personalized experience. The coming months will be critical for outlets to move from Stage 2 adoption to Stage 5 reorganization.
We invite you to share your thoughts on how your newsroom is navigating the AI transition in the comments below.
