We need to bridge the fault line emerging in debates about AI and the future of news

If there’s any place and time when it’s possible to take the pulse of global journalism at a single moment, it’s Perugia in April. The International Journalism Festival, now almost two decades old, draws CEOs and editors, academics and students, institutions and start-ups, regulators and platforms, and many other actors. 

Its hundreds of sessions and panels, and thousands of informal meetings represent a true cross-section of journalism as a business, a social mission, a craft and an identity. Well-lubricated by Aperol Spritz and by the heady hilltop air of an ancient university city, participants tend to speak their minds. There is nothing like it. 

At this year’s festival the pulse of AI in journalism felt bizarrely disjointed. On one hand there were panels about every aspect of applying AI to news, from the use of AI in the largest global newsrooms to the challenges of increasing AI adoption in the newsrooms of the Global South to different approaches to AI news products to tutorials about the use of individual AI tools, all at least partly paid for by Google and by a new US-based AI sponsor. 

On the other hand, there were panel after panel that angrily criticised AI’s potential impact on journalism with some of the harshest and most viscerally emotional statements I’ve ever heard at any conference. One panel even exhibited this disjointedness among participants throughout their 50-minute discussion to the delight of the audience. It was like watching two completely separate conversations that had somehow been randomly intermingled. 

This collective cognitive dissonance can be appreciated by comparing direct quotes from panelists in each of these factions, which I will loosely categorise as the ‘AI Vanguard’ and ‘AI Rearguard’, adapting two categories Rasmus Nielsen used in this piece.  

The ‘AI Vanguard’, of which I am a card-carrying member, generally accepts that AI is happening, that journalism must adapt to this emerging reality, and that journalists and publishers have agency in deciding how they use AI to serve audiences and society. Some representative quotes from this perspective include: 

• “I believe that we can still write the future together. I don’t believe that this is just happening to us. I think we have agency, and we have to just deal with it very actively”. (From Ezra Eeman’s Wayfinder Media Trends Report: navigating tomorrow) 

• “The world is going to change, whether we like it or not” and “Audiences are going to move towards AI. They don’t care to some extent whether the information is that accurate or not – they’re going to do it because it’s easier. We have to get there first.” (From a panel titled ‘Journalism in the age of AI: agents, answers and accuracy’) 

• “AI as a technology is not bad or good – it’s developed by humans. It’s the way in which we decide to use it that makes an impact” (From a panel titled ‘Shaping the future: product strategy for purposeful AI adoption in newsrooms’) 

The ‘AI Rearguard’, in contrast, is clear about technology generally and about AI specifically: it is an enemy and it represents, at best, a gambit by a small group of titans to grasp power and use it to exploit ordinary people for their own selfish gain or, at worse, a cult-like ideology aimed at the literal destruction of humanity, or at least of journalism. 

The Rearguard presence in Perugia was sometimes characterised by voices conveying unconcealed disdain for ‘tech’ and earnest pleas to recognise it as the vile wickedness that it clearly is. Some illustrative quotes from this perspective include: 

• “They’re all mad, and we should actually show people that they’re mad” and “Stop being polite about the whole goddamned thing, because they’re crazy” (From a panel titled “Captured: how Silicon Valley's AI emperors are reshaping reality”) 

• “Let’s be clear. We now have a techno-fascist regime” and “These tech titans are anti-human. They are actually looking to replace humans. They do not see the value of Earth” (From a panel titled ‘The Muskification of American media’) 

• “These companies actually do not want journalists to exist, so if you’re talking about a pragmatic partnership with corporations that genuinely want to obliterate our industry and think that the world will be better off for it, is there really a pragmatic way of engaging in that kind of partnership?” (From a panel titled “After the hype: what are the key trends in journalism and AI”) 

Much of the collective discussion in a gathering like Perugia occurs not just in the public panels but also in the innumerable private conversations over coffee, drinks or a meal. I was, of course, exposed to only the tiniest portion of these conversations, but here too I observed a disconnect. In my minute and highly biased sample I got the distinct impression that many people were, very roughly, moderate members of the AI Rearguard in public but also members of the AI Vanguard in private. This seemed to me to be especially true of senior leaders in news organisations, who were often willing to entertain quite radical ideas about how AI might plausibly disrupt their businesses in the future, but who were also clearly deeply concerned about maintaining calm within their organisations in the present. This certainly seems like a sensible approach to me. 

So what are we to make of all this? 

My own view is that we may be at a tipping point in journalism and in the social sciences that study it that is as much cultural as it is technical. AI has directly threatened the primary activity of these fields – the production of language, knowledge, analysis and ideas – at precisely the same moment as they have lost cultural and political influence as a consequence of wider social and political trends. 

The word ‘existential’ is thrown about quite loosely these days, but I think that many journalists and academics at journalism’s conceptual leading edge might, for the first time, be feeling a fundamental threat to the existence of their profession. They may feel that AI has come for them, and so their reaction (and therefore their language) is visceral and raw. 

This loss of cultural and political influence is probably most dramatic in America, and so it is perhaps not surprising that the most impassioned criticisms of (mostly American) technology companies in Perugia were from Americans. 

The healthcare analogy

Another interpretation was offered by Richard Socher, the founder and CEO of You.com – an AI vendor and major sponsor of the Perugia festival. In a panel titled Journalism in the age of AI, Socher made a distinction between those who are genuinely vested in the outcomes of key societal functions and those who might be vested in existing processes and roles, and perhaps in the pay checks and status that goes with them. 

Using healthcare as an example, he described how the former will likely embrace AI because it can help them to realise those better outcomes, while the latter will likely oppose AI because it threatens the comfort of their existing situation. He is clearly not a neutral observer here, but it’s hard to argue with his suggestion that the future of journalism in an AI-mediated information ecosystem should orient towards positive outcomes for society and audiences rather than towards maintaining an increasingly unsatisfactory status quo at all costs. 

Behind all this, of course, is the weird reality of AI in 2025 – the relentless improvements in benchmarks, the unprecedented adoption rates, the near-consensus of experts about impending AGI, and the direct personal experience available to anyone who spends an hour or two with the latest models and tools. 

Attempts to dismiss these developments as ‘hype’, or to claim that AI just doesn’t work, or to assume that regulatory or legal constraints will somehow turn back the clock, have failed to convince. The claim that “AI will never do X” has usually not aged well, and the implications of that are sinking in. 

I am not the only person to make these observations about the festival, and this rupture in journalism around AI (it is not yet a debate) seems likely to be expressed in other venues in the coming months. The WAN-IFRA Congress in Krakow in May features a panel titled ‘Elevating Journalism to its Rightful Position’, while the INMA Congress in New York will host a panel of CEOs discussing ‘Rethinking the news media ecosystem for a new era’. These are two completely separate conversations about AI in journalism, with very little overlap in objectives, content or tone. 

A way forward

So where do we go from here? 

First of all, we need this to become a real debate, and not merely two disconnected conversations talking past each other. 

The Vanguard (myself included) should probably provide better arguments for how AI adoption can lead to a meaningfully improved information ecosystem, including via actual operating news products and newsroom analytics. The Rearguard should probably communicate with a little less heat and a little more light – with more evidence for their concerns beyond a few cherry-picked anecdotes, and with more carefully constructed arguments. 

A session in Perugia titled ‘How to save journalism from big tech’ provides an example of this approach from an impassioned Rearguard perspective. The age old ‘point-counterpoint’ form of debate is still hard to beat. 

Secondly, we need more proposals of specific and credible alternatives. The Vanguard (myself included) should probably provide more thoughtful and comprehensive descriptions of how AI could deliver positive outcomes for individuals and for society, and of the specific steps we must take to realise those. The Rearguard should perhaps rely less on vague ‘wish lists’ that may be politically and economically infeasible or that essentially amount to a desire to maintain the status quo, and instead provide plausible policy prescriptions or feasible business considerations for decision-makers to assess within the technical, cultural and political context in which all this is actually taking place. 

Specificity has a way of focusing a conversation towards useful criticism and mutually acceptable analysis, and we should all strive for it more often. 

Thirdly we need far more focus on audiences – on ordinary people with everyday information wants and needs. Audiences seem to be adopting AI at an unprecedented rate, for a wide variety of purposes. 

The Vanguard should be more focused on the reality of this adoption rather than on building exciting AI ‘shiny things’ for their own sake. The Rearguard should perhaps acknowledge that audiences are intentionally choosing to use AI, are getting real value from it, and are mostly not being manipulated or entrapped. It is audiences and users who will ultimately decide how AI unfolds in the information flows of our society, based on the value that they get from it, and we should acknowledge their agency in that decision. 

Lastly, we perhaps need more respect and understanding, especially from those (like myself) who may find themselves on the beneficial side of AI-induced change. I have spent a long time working with AI applied to journalism, and yet I can barely grasp the magnitude of what is happening. 

We in the Vanguard should empathise with those in the Rearguard who might have only begun to take this technology seriously with the arrival of ChatGPT and whose disorientation and frustration about what is happening is perfectly natural. Likewise, those in the Rearguard should perhaps not jump to conclusions about the motivations of those of us who genuinely feel that AI can make our information ecosystem function better for audiences.  

I also moderated my own panel in Perugia, titled ‘News or noise? The competing visions for journalism in an AI-mediated society’, with three very thoughtful, informed and engaged experts on the fundamentals of media. We sought to consider the quite radical ideas that Elon Musk has for the future of news on their merits, and to offer criticism, analysis and context in a serious and respectful tone. I hope that we modelled at least some of the above recommendations in that discussion, in spite of the obvious controversy around Musk. That was one of our primary objectives for the panel. 

We are all trying to adapt to ferocious and unsettling change, we all want an information ecosystem grounded in facts and rationality, and we all want to empower and enable citizens in the face of the AI-driven transformation ahead. Surely both Vanguard and Rearguard can agree on that. 

More from us on AI

• Catch up with our conference on AI and the future of news
• Explore our research and original reporting on this issue
• Read our report on how audiences think and it

David Caswell is a regular speaker at our leadership programmes. This article represents the views of the author, not the view of the Reuters Institute.