We are told, frequently, that AI will utterly transform our lives and societies in the coming years. This disruption will be particularly intense, it is claimed, in knowledge-producing sectors like academia, publishing and, of course, journalism. This isn’t just idle speculation, or the pronouncement of experts, but also the common conclusion of most people who engage significantly with AI interfaces like ChatGPT. When you have personally experienced a Large Language Model reading, writing and speaking, articulately and fluently, with encyclopedic knowledge and careful reasoning, then it is quite natural to assume that societal functions based on those skills will be transformed beyond recognition.
But transformed into what?
What would a society whose information flows were fully mediated by AI look like? More specifically, what would the experience of living in such a society be like? How would it differ from today’s familiar media experiences? What would we know? How would it feel?
We don’t know, of course, but it is interesting to observe how difficult it is to even speculate about radically different models of news and societal information. Science fiction has been notoriously uninventive when it comes to journalism, and we lack an array of imagined possibilities commensurate with the flying cars, limitless fusion energy, starships and humanoid robots that drive investment in the physical world. We don’t get to see how the Heptapod aliens of the movie ‘Arrival’ do journalism on their home planet, or how the employees of Starfleet Command receive their news in the San Francisco of the 23rd century. Perhaps news and informational narratives are just too close to us, too deeply embedded in our minds and lives, for us to easily imagine how they might be fundamentally different. Perhaps deeper and broader demand for journalism is somehow constrained by the limitations of its supply.
The speculation about possible futures for journalism that we see most frequently from the news industry and from academia tends to be both vague and pessimistic – a dystopian world of ubiquitous AI-generated misinformation marked by widespread ignorance, riven by conspiracy-driven passions and filled with conflict actively stoked by authoritarians and eagerly exploited by greedy corporations and their AI accomplices. A just deserts, perhaps, for a world that has displaced traditional journalism from its ‘rightful position’ and turned its back on the sober guidance of important editors.
It seems to me that this is a vision of the future of journalism that is oriented primarily by the past – a vision for which the true ideal, the desired North Star, is The Way Things Used To Be. Furthermore, it seems to me that this is a vision that is subtly but significantly influenced by the role of journalists and editors in that future. It privileges, or centers, the importance of journalists in the information ecosystem over the wants and needs of citizens, customers and societies – a warning about the awful consequences of embracing alternatives to the status quo. It is also, in my opinion, deeply infused with paternalistic assumptions about the ability of people to accurately value and effectively consume societal information in ways that benefit them – assumptions that are generally not well-supported by evidence. Lastly, it is a vision that is inherently un-actionable, even futile, because it fights against the largest tides of culture, business and technology.
In this essay I will try to imagine a positive future for news in a world of ubiquitous AI. I will try to be relentlessly ‘user-centric’, privileging individual consumers in their lived lives over the situation of information producers. I will aim to de-center the past and orient instead towards the ideal of a well-informed and high-functioning society in which AI-enabled access to information is an engine that drives individual wellbeing and expands the common good. I will be optimistic not only because I choose to position myself on the AI Vanguard, but more importantly because optimism permits agency in a future that is still unwritten. Finally, I will try to be specific, in the hope that specificity can encourage practical advancement towards positive outcomes, as well as inviting useful criticism and proposals of alternatives.
This essay is the first of a series. In a second essay I will examine investment in a positive vision of AI-mediated journalistic information, making the case that fully embracing the potential of AI can plausibly convert ‘news’ from a damaged sector in terminal decline into a thriving sector exhibiting explosive growth and justifying tech-like investment and valuations. In a third essay I will review specific opportunities for significant investment in the emerging AI-mediated information ecosystem, focusing particularly on ‘line-of-sight’ from existing and likely near-term AI capabilities to the broad diffusion of AI-mediated information throughout society. A fourth essay will focus specifically on culture as both a constraint and increasingly as an accelerant towards a positive vision of an AI-mediated information ecosystem, including at the level of teams, organizations and especially of societies.
These four essays together are intended to form a sort of manifesto for an abundance agenda for news. My primary objective is to inspire ambition. I am trying to make the case that in the long span of human history the application of journalistic information to the cause of a better society is a journey that has barely begun. AI is the next, and perhaps greatest, step on that journey.
How far have we come?
Imagine that you live in Europe in 1425 – perhaps as a peasant on a farm or as a shopkeeper in a small town. Now imagine how much you know about your world. You probably know much about your immediate surroundings, about the activities of your neighbours and about the daily and seasonal patterns of nature. You might learn something from time to time about what is happening in your region from passing travellers, or during your infrequent visits to a market town. You are probably sometimes required to listen to information from monarchs or nobles, delivered by their soldiers, or from popes and bishops, delivered by their priests. And that is probably all that you know. Your ignorance about the world beyond what you personally experience would be almost total. Not only do you not know, but you don’t even know that you don’t know. Your ignorance, and that of your fellow citizens, is likely also reflected in the quality of your life and of your society. It would probably not be a life or society in which a person from 2025 would voluntarily choose to live.
A person in 2025 knows far more about their world than a person in 1425. From the moment we wake in the morning until the moment we fall asleep at night we are saturated with journalistic information. We consume information from a bewildering array of devices, alerts, apps, websites, social media feeds, video platforms and podcasts and, for a few, from printed newspapers, magazines and broadcast news programmes.
We hear that the pope has died 30 minutes after his death. We read about a new study suggesting that red wine is bad for your health. We watch Prime Ministers and Presidents respond to questions posed to them about contentious issues just a few hour ago. We get a feel for a day in the life of a solider in a grinding war on the other side of the continent. We know what Billie Eilish said about Taylor Swift at a party in Santa Monica yesterday evening. We know that an ongoing heatwave in Pakistan is unprecedented in recorded history and is most probably caused by changes in our climate. We hear about a court case in America that has blocked new tariffs from coming into effect. We know that a member of our local council has called for an external audit of a street re-vitalisation project. We know what experts have recently said about affordable housing, or about the prospects for the economy, or about the best washing machine to buy, or about a quick and easy recipe for baked salmon. We know that a football team in Italy is winning a match that began just 30 minutes ago, and that a tennis player in Australia feels confident about a qualifying match she will play in a few weeks.
Here in 2025, we know inconceivably more about what is happening in our world than a person in 1425. Even those of us who are uninterested in news, or who actively avoid news, are steeped in knowledge about events and society that sometimes seems to seep into our minds through digital osmosis.
This explosion of journalistic knowledge has improved our lives to a degree that would be utterly unimaginable to a person of 1425. It enables us to grow as aware and engaged individuals. It provides us with opportunities, choices and control over our lives. It fosters participation, collaboration and accountability. It enables democracy, human rights and the rule of law, as well as markets, contracts and trade. It lets us feel informed. It has, directly and indirectly, played an enormous role in delivering us comforts and security unprecedented in human history, including the banishment of hunger, and the doubling of our lifespans. Most people would eagerly choose to live a modest life in 2025 over even the most privileged life of 1425.
This truly dramatic improvement in our ‘journalistic knowledge’ was driven, obviously, by innovation. Humanity, collectively, figured out: how to print information on paper at scale; how to organise the production of accurate information using the scientific method, journalism and similar truth-seeking processes; how to communicate centralized information using postal services, radio and television; and how to share distributed information using the internet and social media. We have come a very long way since 1425, but we still share a very important characteristic about journalistic knowledge with the people of that time: We still don’t know what we don’t know.
Why would we believe that this centuries-long journey towards more and more knowledge about our world and society has somehow reached its full conclusion? Why would we believe that we no longer have enormous opportunities to improve our lives and our societies by better informing ourselves? What evidence do we have that the current level of activity and service in journalism and news production just happens to be at some kind of end state, beyond which there is nothing more to do? And why would we think this at the dawn of the era of AI and quite possibly of AGI or even ‘super-intelligence’?
We have clearly not reached any kind of ‘end of history’ in journalism, and so we should take the prospect of radically expanding our access to journalistic knowledge in the AI era very seriously. To understand what such an expansion might look like, let’s first consider the extreme limits of manual journalism, from an idealised consumer perspective.
Life as a billionaire news junkie
Imagine you are a news junkie. You really like to know what’s going on. You subscribe to three national or global news sources, plus a local news source. You spend several hours each day reading news on your phone, or on your laptop. You have a list of magazine sites, specialty publications and SubStacks that you visit often. You are informed and knowledgeable – the kind of person that journalists often think of as their audience and that journalists often are themselves. You genuinely enjoy consuming news.
Now, imagine that you win several billion US dollars in a lottery, even though you don’t remember buying a ticket. Your first step in your new life as a billionaire is to hire a chief-of-staff, who quickly assembles a ‘family office’ of highly skilled professionals focused on financial management, philanthropy and legal services. None of this excites you, however, and so you cast about for some way to deploy your new resources on some meaningful pursuit. And then it hits you – you will create a personal newsroom, working just for you.
Working with your chief-of-staff and some expensive consultants, you begin. You first hire a small team of reporters, whose job it is to dig deeply into topics and stories that interest you. One reporter is tasked with spending all day on X, BlueSky, Facebook, Instagram, TikTok, YouTube and LinkedIn, searching out stories that match your interests. Another focuses on government departments whose portfolios interest you, reading their reports, listening to speeches from their ministers and analysing their data. A third spends their time focused on the small town where you grew up, keeping tabs on local gossip, activities and events in ways that help you maintain a connection with your past. A fourth acts as a sort of roving investigator, assigned opportunistically to whatever interests you that day. A fifth and sixth support the others by helping with tasks like fact-checking, interviewing, research and writing.
The new team works hard, and your life improves noticeably. You know more of what you want to know, when you want to know it! But after a while you begin to feel a little overwhelmed. You tire of coming up with stories for your reporters to explore, you are made uneasy by the angles and tone of some of your reporters, and you don’t have time to read all the PDFs and email bulletins that they send you each day. You consult with your chief-of-staff and arrive at a solution – hire some editors!
First you hire a personal assignment editor, tasked with maintaining an awareness of your whims, interests, needs and wants, and then passing that on to the reporting team in the form of story assignments. Next you hire a curation editor, tasked with sifting through all the stories that the reporters produce, and prioritizing which ones you see according to your mood and available time, and a managing editor to oversee the operation of your newsroom. Finally you hire an Editor-in-Chief, tasked primarily with ensuring that the news produced by your newsroom is always accurate, well-verified and trustworthy. After some initial familiarisation, your new editors begin to understand you and to ‘get’ what you look for in news, and the quality of the news bulletins you receive improves significantly.
Now you’re really excited! You huddle with your chief-of-staff and your editors and brainstorm about what to do next. The result is an ambitious plan, which your team begins to execute immediately. You hire some political reporters, tasked with preparing you for the upcoming elections and informing your political donations. You hire some health reporters, who dig into news relating to your various ailments and suggest actions you might take. Soon a nutritionist and a food reporter join the team, working with your personal chef to make sure you’re fully informed about your diet and culinary opportunities. A pair of local reporters join next, keeping you up to date on issues relating to construction permits for your new mansion and initiatives to keep your city clean and safe.
You hire some culture reporters, tasked to continually find and review concerts, exhibitions, plays, books, lectures and courses that match your interests and that would help you to grow your cultural capital. You add a community reporter who keeps up on what your friends and neighbours are doing, and some technology reporters who can advise you on new apps, devices and innovations. You add an education reporter, tasked with bringing you the knowledge you need to give your kids, and yourself, the best education possible. Next come a small science team, a space reporter, a business desk and a trio of economics reporters who examine your investments and inform you about risks and opportunities. By now you’ve also had to hire more editors to prioritize and process all the information that your growing newsroom is producing.
You continue on. An environmental team, a wellness and lifestyle team, a personal meteorologist and an eccentric old lady who generates horoscopes just for you. You hire foreign correspondents, who you place in countries and cities where you vacation or have a business interest. You hire some biographers, who research the backgrounds of public people you are interested in. You hire an ethics reporter, who reaches out to philosophers and religious figures with your questions and moral challenges. Finally you hire some retired football players and some golf journalists to keep you up to date with what’s happening in your favourite sports.
By now you have around 200 reporters and editors working just for you, plus dozens of freelancers who do on-the-ground reporting as needed and a highly skilled team of script writers, illustrators, audio producers and video producers who turn your reports and bulletins into enjoyable experiences better suited to your morning gym sessions or your late nights in front of the TV. After several months of feedback your assignment editor and your curation editor develop a deep appreciation for your tastes and demands. Every person in your newsroom is highly motivated, works long hours and follows instructions carefully, but is also creative, resourceful, thoughtful and deeply loyal to you. Each of them is devoted to providing you with complete information about your life and world, according to your stated wishes, and to directly supporting your day-to-day wellbeing, your civic responsibilities and your long-term personal growth. You begin to feel at ease with your newly enhanced awareness of your world. You learn how to use it and improve it. You feel almost omniscient.
It’s relatively easy to imagine a future scenario in which a well-orchestrated system of interacting AI agents delivers an experience of news similar to this to hundreds of millions of people simultaneously. By stretching the experience of manually-produced news ‘to the limit’ we have gained some sense of what the scaled automation of news gathering and news production might lead to. We all might soon have our own personal newsrooms.
But this is grossly insufficient.
The application of AI to journalism offers far more than merely the automation, and thus democratisation, of the news experiences of a billionaire, a Prime Minister or a President. AI gives us capabilities that are simply not available from any newsroom staffed by humans – regardless of size or budget.
When we consider what, exactly, the personal newsroom provides us that we don’t already get from our more familiar news consumption experiences, we find that there are two broad capabilities that are fundamentally different: The ability to do personally relevant journalistic tasks at greatly increased scale; and the ability of the consumer to easily interact with and control the journalists doing those tasks.
The billionaire news junkie story took those two fundamental capabilities ‘to the limit’ as a manual process, and we can use the same technique to get a sense of what might become possible from a newsroom built around AI.
Infinite news
What would news at AI scale look like ‘in the limit’?
For a large and expanding portion of journalism, AI lets us abandon the assumption of scarcity in news gathering and reporting, and with it the notion of selective ‘newsworthiness’. When it comes to news from digitally accessible sources – a large and increasing portion of news – we will soon be able to build an essentially unlimited editorial workforce, including personal editors and curators of stories and experiences. To understand how this might affect our experience of news as individual consumers, let’s consider how this might play out within a single geographic news market.
Imagine the news stories currently generated each day in your own city or local region – perhaps from a few metro dailies, a few TV newsrooms, a few dozen news websites, a handful of magazines and some blogs and TikTok channels. This might amount to, generously, a few thousand stories per day, each deemed to be ‘newsworthy’ by some journalist or editor. Many of these stories are lightweight lifestyle or culture stories, or personal ‘takes’ by journalists, or accounts of something that has happened, or simple transfers of information from a source to a story, etc. Some small number are deeply investigated and comprehensively reported.
Now, imagine the possible stories, each relevant to at least one person, that might ‘naturally’ exist in that same area in a single day if sources for them were available and reporting on them was feasible. In a city of two million adults, using a generous estimate of, say, 20 consumable local stories per person per day and allowing for, say, 95% overlap, this might easily amount to around 5 million ‘natural’ stories per day – many of them unique to a single individual. Manual journalism therefore currently covers, in aggregate, something like 0.05% of the conceivable relevant local or regional stories. These numbers, definitions and assumptions are highly suspect, of course, but in the absence of useful empirical research these approximate orders of magnitude are quite plausible.
Now imagine the effect that this very familiar, but deeply impoverished, trickle of stories has on the relevance of any one individual’s news consumption experience (defining ‘relevance’ as ‘actually mattering to the individual’). Even assuming that you could somehow access the 10 or 20 most personally relevant stories currently produced in your city each day, across all publications, you are still drawing from a pool that is almost empty.
AI can fill that pool. By reading, watching, listening to, analysing and integrating any digitally accessible news source – including via email or online interviews, contracted freelancers and sensors – AI could increase the percentage of ‘natural’ stories covered from maybe 0.05% to perhaps 1%, or 10%, or perhaps even 50%. Every story that could be covered, would be covered, even if only relevant to a single individual.
This news coverage at AI scale would be substantially different from sporadic manually-selected news coverage, and not only because of the increase in the volume of stories, or because of the improvement in the relevance of stories from the perspective of individual consumers.
The systematic and methodical nature of the coverage would change what people could expect from news – you would just ‘expect’ that everything that could be covered would be covered, and you could act on those expectations. What were all those police sirens I heard on my way to work this morning? Does my council spend more on street sweeping in wealthier areas? Is that volcano in Sicily still erupting? What are my upcoming opportunities to voice my opposition to my government’s policy on some issue? The expectation of knowing would probably change – expand – what you desired to know.
The removal of scarcity from newsgathering would also substantially expand what could be considered ‘news’. Stories about everything your company’s CEO does, or about when flowers are blooming in a nearby park, or about legislation affecting building supplies, or about activities for mothers of 2-year-olds within strolling distance, or about the blogging history of a politician all can become ‘news’, because reporting on them is now feasible.
And this expansion of the definition of news would be open-ended. Consumers could continually add new topics or subjects to newsgathering as they developed interests or as they began to appreciate the new power available from AI newsgathering. News would evolve rapidly, driven by consumers. News coverage would grow and grow in scope as well as in scale, eventually creating a truly ‘infinite newsroom’, constrained only by the wants and needs of consumers and the capabilities of AI reporting.
This concept of ‘infinite news’ is, frankly, horrifying to many people.
Feeling overwhelmed by news is something that most of us already experience, sometimes to the point of feeling captured by compulsive scrolling, shattered by fractured narratives and addicted to our devices and platforms. Increasing ‘news avoidance’, especially by younger audiences, is at least partly driven by this sense of overload and fatigue. Surely ‘infinite news’ is likely to make us more overwhelmed and confused rather than more enlightened and empowered?
If we, as individuals and as societies, are to reap the benefits of AI’s removal of ‘supply side’ constraints on news, how might AI also help us with these ‘demand side’ constraints? How might AI help us to improve our relationship with news?
Comfortable news
What would consumer control of AI-mediated news look like ‘in the limit’?
The most obvious need of any consumer with access to ‘infinite news’ is curation. In the current information ecosystem we essentially have only two ways of curating digital news: manual curation of stories by human editors, and algorithmic curation of stories by ‘recommender systems’, using machine learning to predict preferences based on previous clicks and engagement.
Neither of these is capable of personal curation in an environment of open-ended or ‘infinite’ news, because both curate only the artifacts of journalistic information – articles, videos or posts – rather than the information itself. Manual curation of one-size-fits-all articles cannot scale and cannot truly personalise. Recommender systems can only recommend artifacts based on thin signals oriented towards our immediate whims and instincts and ignore our thoughtfully considered information preferences. Both are insufficient.
AI enables a third kind of curation – a personal, but infinitely scalable editor. Like the billionaire’s human editors, this is an editor that we can speak to, instruct, guide and correct and that will ‘get to know us’ over time by remembering and accumulating our personal preferences. But unlike the billionaire’s human editors this personal AI editor is quite comfortable with ‘infinite’ and open-ended news. These kinds of ‘personal editors’ are already emerging, for example Spotify’s ‘AI Playlist’, which allows users to generate personalized playlists based on text prompts. Personal AI editors might give us a power over our news that we have never experienced before – a new power that might feel to us like opening a newspaper for the first time in the 17th century would have felt.
Another obvious need of a consumer with access to ‘infinite news’ is the ability to consume information with far less effort. When we feel ‘overwhelmed’ by information it is often not the volume of information that is the problem, but the volume of information artifacts that we must process. We are overwhelmed by the cognitive labour required to: decide which artifacts to select; to identify and isolate the new information contained within them; and then to contextualise that information within our existing knowledge. This is very hard work.
In the existing information ecosystem we have essentially insisted that news consumers adapt to our format – usually the one-size-fits-all text article – because we have been unable to adapt our news to consumers. We don’t really know how news consumers – either individually or en masse – would choose to experience news in the absence of those constraints, but the meteoric rise of video and audio news consumption over the past decade suggests that they are eager for alternatives.
Early use of AI by newsrooms has focused on re-versioning individual artifacts – turning articles into summaries, turning audio files into draft articles, etc. – with the goal of expanding the accessibility of news. AI will eventually enable us to abandon canonical artifacts entirely and focus instead on communicating stories and information directly to individual consumers in the most effortless way possible. We already see early examples of this emerging, such as the ‘catch-up’ feature offered by Verdens Gang AS in Norway, which builds a personal summary of developments in a story since a consumer’s last visit, using event descriptions rather than articles. These early efforts are, obviously, the first small steps of an inexorable journey towards more and more adaption of news to consumers – a journey with no logical ending point short of effortless consumption.
A third obvious need for a consumer with access to ‘infinite news’ is the ability to interact with that news – to essentially talk to, or text with, or otherwise engage with an accessible expert with encyclopaedic knowledge of the subject or story. This is the kind of consumer control of news that is easiest for us to imagine, because we are already familiar with the experience of interacting with ChatGPT and its kin. But our relationship with chatbots is still in its earliest infancy, and the use of chatbots for consuming news is likely to evolve considerably in the coming years.
An example of this evolution is ‘agentic’ AI – systems that can act autonomously in pursuit of high-level goals, that have persistent long-term memory and that can access external tools like APIs and payment systems. AI agents clearly have many uses in the gathering and processing of news, however their impact in the experience of consuming news is likely to also be significant – and not merely by enabling us to explore stories more thoroughly through conversation. Autonomous agents can, for example, act asynchronously or even continuously. They can access large volumes of ‘news’ and use it to act on our behalf, without us ever needing to know the details. They can do investigations for us, they can monitor something or someone for us, they can use ‘news’ to negotiate for us – perhaps with other agents. Agents open up the potential for us to ‘delegate’ the consumption of entire categories of news and then to use that news to ‘just take care of it’ for us, whatever it is we need, or to ‘represent’ us by adding our ‘voice’ (or even vote) to political and cultural conversations. In other words agents might eventually enable us to consume news and to act on news without using our scarce personal attention at all.
Another need for a consumer with access to ‘infinite news’ is an ability to optimize between ‘lean forward’ and ‘lean back’ experiences of news. ‘Lean forward’ experiences are those which require active decision-making by the consumer – such as the formulation of questions or instructions – whereas ‘lean back’ experiences are those which enable passive consumption of an experience that has been constructed for you by a producer. ‘Lean forward’ is asking ChatGPT a question, whereas ‘lean back’ is listening to a podcast. ‘Lean forward’ requires effort, initiative and energy, whereas ‘lean back’ does not. AI enables us to produce personal ‘lean-back’ experiences of news on demand.
The ability to consume news as lean-back entertainment is already well-established in our pre-AI information ecosystem, in genres such as news-related comedy or ‘true story’ fictionalisations. Extending and personalising this technique using LLMs, video models, multi-modal models and agents as AI media producers opens up new ways to experience personally relevant news with significantly less cognitive effort and friction. When combined with personalised curation from personal AI editors and personal control over formats, context and styles this capability might make the consumption of news significantly more comfortable for many consumers.
Finally, a primary need for a consumer with access to ‘infinite news’ is the ability to trust the accuracy, completeness and authenticity of that news – the need to feel safe in one’s news consumption, without risk from deception, manipulation or incompetence. Satisfying this need is primarily a matter of the design of newsgathering and news productions processes and systems but also requires consumer-facing ‘proof’ of trustworthiness.
Applying AI to trust in news is very challenging even to define, much less execute, however an excellent description of the problem is provided in Yuval Noah Harari’s recent book ‘Nexus: A Brief History of Information Networks from the Stone Age to AI’. Harari describes how ‘truth-seeking’ and ‘self-correcting’ mechanisms like the scientific method or journalism itself were crucially important for humanity to take advantage of the huge volume of information unleashed by the printing press. He lays out the core challenge as finding a similar mechanism for the era of AI-mediated information.
There is every reason to believe that, with sufficient effort and focus, a robust and dependable mechanism for orienting information towards truthfulness in an AI-mediated ecosystem can be developed. Such a mechanism would not be perfect, just as its human editorial predecessors were not perfect, but it could be worthy of our trust. We already see significant experimentation in related subjects like automated fact-checking, bridging algorithms for content moderation, cryptographic authentication of media and even around emergent truth-seeking behaviours of language models. Becoming authentically comfortable with the truthfulness and trustworthiness of ‘infinite’ AI-mediated news will require significant innovation, perhaps including regulatory innovation, but this is a challenge that we have only just begun to address.
Eventually, therefore, AI might bring us an informational experience of our world and our society that is vastly broader and deeper than the news we experience now, while simultaneously being easily and effortlessly accessible. News that is both ‘infinite’ and ‘comfortable’ might conceivably place us in a relationship with news that is as far advanced from our current experience as our current experience is from that of a news consumer in 1425. Furthermore, by empowering – or ‘super-empowering’ – us with information, and by rendering society far more transparent and understandable, AI-mediated news might unlock dramatic positive outcomes. It might potentially usher in a far more prosperous, fairer and saner era in which everybody has an opportunity to flourish and thrive.
But what, specifically, would this information utopia – this ideal, this perfect world of news and information – look like? What might we know in such a world? How might our lived experience of it feel?
What might we know?
We could know (or at least have an opportunity to effortlessly know) much, much more in terms of breadth and depth of information, but we could also know that in ways that centred entirely on personal context – what it means for us, what it does for us, how it connects to us, how it helps us or hinders us and how it satisfies us.
The word ‘personalisation’ has lately become somewhat tired – often used in a lazy way to refer to ordinary recommender systems offering ‘one-size-fits-all’ content or as some vague placeholder for some kind of computational future of news. But ‘personalisation’ can be tangible, vital and palpably felt. It can mean an information environment that is fully, unapologetically and unambiguously centred on the individual and their wants, needs, situation and context. It can mean an experience of news in which the person and their context is deeply integrated into every experience, every story, every emphasis and every choice. It can mean an information environment in which we almost never consume ‘one-size-fits-all’ news content.
To get a feel for what, specifically, we might know in such an information environment, let’s look at examples of radically expanded and radically personalised journalistic knowledge through several different lenses: a geographic lens; a temporal lens; and a topic lens. Keep in mind that while there are obviously significant hurdles to actually implementing news experiences like those portrayed here, as well as many potential harms associated with them, our immediate focus is on imagination and potential, not on challenges or harms.
So what, exactly, could we know in an ideal news environment?
We could know far more about our world and our experience of it at every distance from us, extending outward from our most personal and private issues to the largest universal issues facing humanity as a global civilization and a species.
We would know more about ourselves: What are the big stories from my home life this week? How is my career going and what skills should I focus on? What should I buy at the farmer’s market on the weekend? Is there any new information, context, analysis or advice about my recreational experiences, or about my health situation, or about my personal finances, or my car or my property, or about my past successes and failures and my future opportunities? Each of us, personally, could be the subject of any personal story that might be relevant to us and for which an AI could access a source, perhaps from our devices, our online accounts, our digital records or our self-reporting.
We would know more about our relationships and our family: What are my friends and family up to this week? Is everyone OK? How can I help a needy friend? Who might we invite to brunch at the weekend? What is my cousin’s new business venture about? How is our old roommate’s move to a new city going? What’s the backstory on that guy we met on Monday? Where did Pedro and Stacy go on holiday this year? Is my kid any good at that video game she’s always playing? All of us could know anything that anyone close to us chose to share publicly, or on social networks, or perhaps in private communication with us.
We would know more about our immediate neighbourhood: What is that building site down the street going to become, and when? What happened with that awful car crash on the corner last week? What’s fun and safe for teenagers to do on weekday evenings this month? Where do new mothers shop for groceries around here? What have the local council been discussing or planning that would affect our street? Who bought that farm out on highway 12? Is our village losing young people? Are more artists moving here? Why is that helicopter hovering over the park? Did they catch that guy who vandalised the library window? How should I use the sports centre? Why did the Thai restaurant close, and who’s opening that new café across the street? What did the new pavement on the next street cost? Was that a good deal? Did any of my neighbours pass away? Did any have a new baby? Can I bring them anything? What else happened nearby that I might find important, practical, interesting or gossipy? We could know almost everything that happened in the environment of our daily lives, as long as AI could access that information.
We would know more about our local government, and about our local representation in government: What does the council collect in revenue from our town? What does it spend here? What does it spend it on? Was that spending effective? Where would I see a nearby example of effective or ineffective spending? How does our council compare with other similar councils in its priorities and effectiveness on issues that interest me? What is my local councillor or my Member of Parliament interested in? What do they spend their time on? What has she said? What has she done? What has she said or done that affects me, personally, or than might affect my neighbourhood? What might another candidate from the last election have done differently? What other local councillors or Members of Parliament are saying things or doing things that might affect me or my neighbourhood? Is my local councillor or MP a kind person? Are they trustworthy? What are their values? Who have they helped? What do other people think of them? What have they done outside of politics? How does the council affect the energy rates or the water rates? What programmes do they have that I could take advantage of? We could have almost complete transparency of public governance and its impact on our community and on ourselves, personally.
We would know more about the other parts of our city or county that we spend time in: What’s happening in the area where I work this week? Are there any roadworks or street fairs or shop sales or evening concerts? How is that area changing? Is it generifying? Are there more homeless people? Are there any interesting neighbourhoods or historical sites within lunch-time walking distance? How do most people there commute from my home area to my work area? Why are the streets so crowded on Saturdays? Are any protests happening? Why here? What is the business community here like? What’s the story with that strange-looking pub we pass on the way to school? Is the air quality different here than at home? Why does this area have less nice shops than my neighbourhood? How safe is the area around the cinema at night? Who is making a real difference helping that homeless community? What unfamiliar part of the city should I explore? Why? We could know far more about other areas that we frequently visit, and about the issues, opportunities, threats there that are relevant to us.
We would know more about the city we lived in: How was I represented in city government this week? Was my neighbourhood or council area mentioned or discussed? Is the city responsible for that bridge where traffic is reduced to one lane, or is that the county? How can I complain about that? Is the city council competent? Which members? Are they respectful of their city’s people, and of each other? Is my city councillor competent? Are there public consultations on issues that matter to me? Where and when? What ‘feel-good’ stories happened in the city today? What kinds of crimes are most common, and how does this compare with other places? Are the police solving them? Is the city investing in infrastructure? What are the projects? Are they within budget and on schedule? Will they help me? When? Who is being left behind in the city? Why? How well are the schools doing? The hospitals? The waste removal? What kinds of businesses are doing well? Why? What unique cultures are expressed in this city and how can I experience them? What makes our city feel like a community? What does the city budget look like? Where does the money come from? Where does it go? What do people in other places think about my city? What did people who visited here like or dislike? We would know anything about our town or city that might be relevant to us, including about those things that give us an identity as part of our larger community.
We would know more about the larger region that we lived in: What major new regional infrastructure is planned? What is the condition of existing infrastructure that we depend on? What about airports? What about energy, and particularly renewable energy? What is blocking or delaying those initiatives? What are the climatic trends in this region, and what are their effects on me and my neighbourhood? How is the water in the region managed? What about floods? How much of the people in the region live in flood zones? Who pays for their insurance? How is the region planning for and dealing with natural disasters? How does agriculture, industry, or tourism drive the regional economy? What is happening to wildlife in our region? How are the lifestyles in other areas of our region different than the lifestyles I’m familiar with? What cultural events or traditions are part of other areas in my region? Who is protesting across the region and what are the controversies that I might support or oppose? How are our resources being managed? How are younger generations being educated and developed across the region? We could know every relevant aspect that helped us to situate ourselves within the geographic, climatic, traditional, economic or cultural context of our region.
We would know more about our country: Did our parliament or our government do anything that might affect me, or that I might be interested in? Did the politicians who represent me say anything, or do anything or go anywhere or meet anyone that I should know about? Did they fail to say something or do something that might go against my interests or my values? What is the mood of other people in our country, from the smallest village to the largest city? Why? What are the patterns and correlations? What do we take pride in together? What are we all concerned about? What do we all agree on and what do we disagree on, and do I agree or disagree on those things? What parts of our country or what people in our country do I know the least about? How could I learn more? Where should I travel to next? What do other people think is important for our country? What do prominent, accomplished or famous people think I should know? What do ordinary people in different regions think I should know? We could continually maintain a visceral ’feel’ for our nation and for our place within it.
We would know more about our continent: How are we, as a community, and me, personally, connected to nearby countries? How do those countries work, politically, economically and culturally? What’s going on there that might interest me or that affects me in some way? What value does my country share with those countries, and how do those fit with my values? Are our politics similar or different, and can I learn something about my own political stance from that? What are their important issues? How do nearby countries support us or oppose us, in politics, in values, in trade, etc? What continental institutions or governance organisations are important to all of us? Which affect me, and how? Do we share any cultures, identities or history, across borders or languages? What things are happening now that affect them in the same way as it affects us? How can we work on those shared challenges together? Who is working to divide us? Who is working to unite us?
We would know more about our world: What is happening across the world that might affect me, interest me, be important to me or that I should know about? What are the global trends of politics, trade, culture, etc that I should pay attention to? What is emerging in places around the world that might have special resonance for me? What is happening to the natural world? Which places are prospering, which are in decline, and why, and what can I learn from that? What scientific discoveries and inventions have happened? What innovations of business or art or culture should I know about? What trends do they reveal? How might they move my life forward or hold my life back? What culture or style from other societies, or in other languages, should I explore? What should I learn? What risks should I take? Where should I explore?
The potential to use AI-mediated information to situate and contextualise ourselves within any granularity of geography is clearly dramatic, however AI can also help us to develop and maintain a significantly deeper understanding of where we are situated in time. This temporal understanding would be manifested as vastly better personal and collective memory of the past, as much greater appreciation for what was happening right now, in real time, and as much more sophisticated and informed sense for what might happen in the future and of what might have happened had things been different.
We would know more about what happened in the past: What did that politician say he was going to do? How has the Prime Minister’s public rhetoric on this subject changed over the last 3 years? How has it changed since his university days? What is the trend on egg prices, or the consistency over time of sentencing in terrorism trials, or the exception to the norm on the emigration rates of young people? What prominent person’s statement is contradictory with the past, or hypocrisy, or a change of mind, or an error or an excuse? With a near-perfect memory of essentially everything, our information might become a deeply contextual blend of historical education, journalism as a ‘first draft of history’, record keeping and something like personal automated journaling.
We would know more what is happening in the present: Has the meeting started? What is it like to be in this war zone, or at this concert, or this protest, or this joyous celebration, at this exact moment, and how can we experience it as it happens? What is the latest thing we know about the situation in this flood-stricken community, or the progress of that life-saving operation? This broad access to real-time monitoring, or even automated live blogging or live streaming, might enable all of us to personally witness, or even participate in, many of the most significant events of our societies.
We would know more about what might happen in the future: What are the range of scenarios for what might happen in the election, and how would each of them affect me? What might affect how ongoing events in a political crisis develop, and how might I influence that? What are the chances of me being affected if a particular economic outcome takes place? How would others be affected? What are other people saying might happen because of the delay in an infrastructure project, especially people who are well-positioned to know? Who is speculating like a person in a pub conversation, and who is making informed, logical arguments?
We would know more about what might have happened: If the other candidate had won the election, what might have happened in parliament regarding the new regional development programme? If the local council had passed the tax increase, what might be different regarding the response to the flooded street in my neighbourhood? What might have happened if I had chosen to travel on this airline for my holiday instead of on that airline?
AI-mediated information could unlock access to deep knowledge about any topic, subject or category of news, independent of geography or time. Furthermore AI might expand our assumptions about what a ‘topic’ or a ‘subject’ or a ‘category’ of news can be, bringing in far more nuance, pragmatism or personal impact than is feasibly for news today.
We would know more about politics (and about our own power in society): Our personal politics, the sources of our politics, our confidence in our politics and the expression of our politics at every level of political competition would be genuinely transparent to us, regardless of our interest in legacy political coverage. We would know about our opportunities to participate politically, about the alliances we could be part of, about the actions we could take and about the latent influence that we might have. We would know more about the positions of others. We would know where other people stand or probably stand on issues, based on their public statements, their proposals, their declared values, etc. We would know more about the interface between politics and policy. We would be able to effortlessly monitor the speech, actions and votes of politicians and politics to assess their fidelity to their pledged positions and their implied positions. We would be able to effortlessly monitor the implementation of policy and the effectiveness of policy including in terms of avoidance, excuses, etc. by politicians. The real mechanics of power, such as the complexities of selection in the UK’s labour party or the procedural rules of the US senate, would be transparent to us. We would know more about our real political options, about who was accountable for what, about what did or didn’t work, and why.
We would know more about economics (and about our contributions & consumption): We would understand our personal economics and the economics of our experienced lives – where does our money come from? Where does it go? Who do we subsidise? Who subsidises us? What are we obligated to provide and entitled to receive? Who can we help and how should we help them? How are we contributing to our household, our neighbourhood, our town or city, our country and our planet? What are we costing each other and our environment, at every level of our lives? We would know about our economic opportunities – about how we might contribute more or cost less to our communities. We would know how we might become more valuable, in our jobs, careers, businesses and our volunteering, and how might we claim or share more of that value? We would know more about the capital we accumulated, not only in savings and investments but also in government obligations, environmental capital, infrastructural capital, intellectual capital, etc.
We would know more about art & culture (and about its potential in our lives): We would know much more about our cultural opportunities, about how those opportunities might provide meaningful personal growth, and about our opportunities to acquire cultural capital. We would feel far more confident in investing in culture, because we would understand far more clearly and specifically what we might gain from that investment. We would more easily understand the culture that we create, participate in and consume, in ways that were personally meaningful to us. We would know when specific cultural contributions that might be particularly valuable to us were created by others – a new point of view described in a novel, a new piece of music that captures a ‘vibe’ that is important to you, a connection between a personal intellectual challenge and a recent cultural or artistic work. We would know more about the economy of culture, about the role of art and culture in status and about the changing cultural ‘climate’ of our communities. We would have a far better ‘feel’ for the totality of human culture across space and time, including its frontier, and we would know more about where we and our experiences fit within that.
We would know more about lifestyle (and about developing or expanding our own lifestyles): We would know far more about the lived reality of our lifestyle options, about how we might choose between those options and about how we could realise our desired lifestyle. We would have more knowledge and control over our lifestyles at the broadest level – Should be become parents? Should we move to the city? Should we orient towards tradition or towards adventure? We would have far more intentionality in the small day-to-day expressions of our lifestyle – what should we eat? How should we dress? Where should we go for a walk? We would know more about how to choose where we live, and how to arrange our home to suit the lifestyle that we wished to experience, and about how we should budget our discretionary time and money. We would know more about where we could travel to, about what we might experience while traveling and about how that experience might change who we are and how we lived. We would know how to eat better – whatever ‘eat better’ meant for us. We would know more about style, about having a style and about whether we should be deliberate and intentional in our personal style. We would be far more informed about the connections between our lifestyle, or aspects of our lifestyle, and our wellbeing, health and prosperity.
We would know more about education (and about our own opportunities for personal growth): We would become far better ‘lifelong learners’, regardless of our personal curiosity, because we would each have unlimited guidance and personal tutoring. We would know more about our opportunities for personal growth through improving our knowledge and understanding – more about what we could become. We would know far more about our opportunities for realising genuine meaning and purpose in our lives, and also about ways in which we might improve our status. We would all understand the potential paths for actual, palpable personal development that were available to us, no matter how intellectually impoverished our beginnings or mediocre our talents. We would each know more about our intellectual possibilities – about what we could be most interested in and intellectually motivated by. We would know about new knowledge or new ideas in fields we were interested in – whether string embroidery or string theory. We would know about the ideas of people we were particularly interested in or respected, and about their new ideas. We could cultivate and follow our curiosity, however strong that might be – launching and pursuing our own investigations, genuinely ‘doing our own research’ while fully supported with context, fact-checking and guidance.
We would know more about our identity (and about shared history and experiences): We would know what we shared with others, and we would have more opportunities to fully experience the communities that we were a part of or wanted to be a part of. Communities of shared history, or geography, or shared experience, or language or religion or loyalty to a sports team or fascination with a movie or an idea would become far more available to us. We could more easily access the shared stories of those communities, and better understand their cultures, expectations and benefits, and how we might participate in them in ways that added to our wellbeing. We would also better understand and navigate the complexities, contradictions and trade-offs in and between our communities – better able to maintain our loyalties and connections while remaining pluralistic or even cosmopolitan. We would, in particular, know far more about our intellectual communities – about who was interested in what we were interested in, about who is thinking in ways that we are thinking, or in ways that challenge us, and about who would appreciate what we can contribute. As we learned more the values, activities, actions, factions, controversies, questions, heroes, villains and vanities of our various communities we would, perhaps, become more selective in the identities we chose to attach ourselves and in how we expressed that attachment.
We would know more about justice (and about safety and fairness in our lives & communities): We would know more about the true threats of injustice towards us – to our physical safety, our financial safety, our psychological safety and the safety of our livelihoods and social standing. We would know about who threatens us and why, about where the most significant risks are, and about how we might minimise those risks. We would better know when we were being ‘ripped off’, or ‘played’ and we would know when we were getting a ‘good deal’ and appreciate it. We would know more about injustice to others – about who was suffering injustice in our street, our city, or nation or our world. We would know more about what we, personally, could do to correct a specific injustice against a specific person, or to genuinely improve the justness of our community or society. We would have a far better idea of how our own values intersected with our assessment of fairness – a clearer view of the personal responsibilities of ourselves and others, and a more accurate sense of what is and isn’t someone’s ‘fault’. We would know more about the processes available for correcting and compensating for injustices, from our social systems and the criminal justice system to civil and commercial claims. We will still have injustice, of course, but our ideal information environment might also give us meaningfully more transparency and therefore meaningfully more resolution of injustices as well as meaningfully more understanding of – and maybe forgiveness of – injustices. To paraphrase Evelyn Waugh, perhaps to know more is to forgive more.
We would know more about ethics (and about our own lived values): We would have a clearer sense of our own values – what they actually are, what they are they based on and how closely we live by them. We would know much more about who or what was offending our values, and about what we could do in defence of our values. We would know who or what was affirming our values, and about what we could do to support that. We would be better able to interpret stories, issues, events and problems in terms of our values, and to understand how they might inform or grow our values. We would know how to be more morally consistent at every level of our lives, understanding how to think about stories, and about people in stories, in ways that reflected who we would like to be, and understanding how to act on stories in ways that reflect our moral intentions. We would also know more about how we might connect with others in a moral or spiritual community, whether locally, nationally or globally. We might, in a very real way, learn how to become the change we would like to see in the world.
We would know more about problems and solutions (and about what we can do): We would know about our own problems and about ways in which we could make our lives better. We would know what each problem means for us – how did it happen? Did I do something to cause the problem? How serious is it? What can I do to resolve it? What help can I get? Who has a similar problem? How did others approach similar problems and what solution worked best? Whether financial, medical, family, social, housing, transport, educational, legal, housing, behavioural, relationship or spiritual we would have the knowledge to better identify and define our problems, to identify plausible solutions to them and to act on those solutions. We would also know more about the problems of our community, including problems of complexity, of expense, of disagreement, of social dynamics, of incompetence, of natural phenomena, of history and of bad luck. We would know more about the structure of those problems, about the real issues at their root, about the costs and hard choices necessary for their resolution, about their effect on us and our contribution to them, about the people working on them and about the best way that we, personally, could help. We would know more about the problems of our country and our world, including problems of conflict, of climate and environment, of economics, of energy, of resources and of national power. Here too the fundamental basis of those problems would be clearer to us. We would better understand their effect on us and our role in them, and we would know the most effective leverage that we, personally, could apply towards their solution.
We would know more opinion (and also more about opinions): We would know more about what other people think. We would know who has fresh and valuable opinions about issues that are most meaningful to us. We would better understand the spectrum of ‘takes’ on stories that were important to us and would know much more about how those takes added new insight or perspective and which aspects of stories remained vague, unexamined and uninterpreted. We would know the essence of each opinion, and about how typical or atypical it was of its author. We would have easy access to opinions which challenge us without enraging us, which support us without pandering to us, which are influential and spreading, and which are exhausted and declining. We would know who shared opinions. We would be better able to understand what the larger communities of social media or cultural institutions were thinking – the view of millions reduced to an ‘op ed on demand’ – while simultaneously be better able to find unknown individuals with a unique interpretation, an unusual depth of analysis or a quirky spin on a familiar issue. We could monitor the view of people we were particularly interested in – perhaps, if we wished, receiving late night alerts that so-and-so has changed her mind about such-and-such. We could know all this in ways that supported us in the formation of our own opinion, and that made us and our views a full participant in the landscape of opinion.
We would know more shocking & scandalous stories (and have more to gossip about): We would constantly have unique and interesting things to talk about with others. We would know outrageous and salacious things that were titillating to share, revelations about celebrities, rumours and inside information about important and self-important people, party talk and pillow talk. Those who enjoy ‘news’ about others, parasocial relationships with celebrities and just a good gossip with friends would have more to talk about than ever before. In an ideal information ecosystem we would simultaneously have access to an endless stream of titbits and scoops, available only to us, and suitable for use as social and conversational currency. Ideally, and perhaps improbably, this would be matched by a respect for the privacy of others, and perhaps by a degree of social or even centralised control over information about private individuals or about the private lives of public people.
We would know more interesting or serendipitous things (and have more fun knowing them): We would know about curious or unusual things that we might enjoy knowing, or that our friends and family might enjoy knowing – unusual events, mysteries, coincidences, comical situations, things that had a particular connection with you, personally, or with a loved one. These stories would not be exploitive, like an endless stream of cat videos on a social media feed, but genuinely useful and valuable to ourselves and to others – even if only in social or even frivolous ways. We would also be exposured, frequently, to issues, stories, facts, interpretations and analysis from outside our usual information diet, all offered to us as an opportunity to engage with those subjects or topics. Ideally this serendipitous exposure to other interests would, over time, provide us with opportunities to find new interests, or to expand or redirect our interests in new directions that might appeal to us. Our ideal experience of serendipity, over the long arc of our information-consuming lives, would be to optimize our interests towards our own personal wants and needs – to eventually expose us to the entire space of topics in ways that let us ‘select’ our own interests and chose the objects of our own curiosity.
We would know more of what we should know (or what others think we should know): Most people hold the opinion that there is news that we may not be interested in, that may not directly improve our lives and that may not serve any purpose to us, but that we nonetheless should know anyway. In the legacy news ecosystem a handful of unknown editors decided ‘what we should know’, based on their assessment of a story’s importance to society, it’s potential to edify the public or on a perceived need for common stories within a community or nation. That ‘we-will-chose-for-you’ system of curation is increasingly unacceptable, however a system based on the voluntary selection of people or AI agents that we explicitly allow to select news for us is quite feasible in an AI-mediated ecosystem. In this approach to curation we would voluntarily choose people, agents, organisations or even governments to whom we were willing to hand over some control over the stories that we were exposed to. We would then know what those people, agents, organisations or governments thought we should know, from a basis of trust and respect rather than of compulsion and paternalism.
We would know more about knowing more (and assess information and perspectives more critically): We would have a profoundly different relationship with journalistic information, because we would know and experience our own informational power. We would have more awareness of difference - of opinions, of interpretations and of narratives. We would receive personal inspiration and motivation from others, and from the relationships that others forged with information in this new ecosystem – learning together how to use our new omniscience. We would be able to better critically assess information and narratives, because we would be empowered to effortlessly compare, contrast, assess, verify and understand. We would allocate our trust more effectively, and because of that we would trust more. We would learn what was truly worth our precious time, and what was not.
Any topic, subject or category of news we could conceive of would be available to us – any micro-beat in any niche that we found relevant, regardless of how obscure or nuanced. Every single morsel of information we consumed would be deeply personal – fully interpreted for us in terms of the real needs and wants of our actual lived lives. We would not just have access to information but also to brilliant analysis producing many ‘takes’ and generating many insights, all oriented towards us and done just for us. We would know far more, and understand far more, in far more useful ways, than even the most privileged news consumer of 2025
How would it feel?
The consumption of news and other societal information in our information utopia would feel easy – far, far easier than it does now. Some news would still, of course, cause us anger, sadness, anxiety or uncertainty, but the way in which we accessed news and our fully encompassed relationship with news would, ideally, feel significantly healthier. Similarly, in this imagined perfect world, the expanded scale and scope of news and the expanded definition of ‘news’ would contribute to a deeply beneficial feeling of awareness, of understanding, of knowing, and perhaps eventually of wisdom.
To get a sense of how, specifically, it might feel, let’s try to deconstruct the feeling of news consumption and look at examples of each contributor to an ideal experience. As with our speculation about what we might know, keep in mind that our immediate aim here is imagination and potential, not on implementation challenges or possible harms.
So what would ideal news consumption feel like?
It would feel like being in control: In an ideal AI-mediated information ecosystem we would feel calm and in control. We would feel far less overwhelmed by information than we often feel now, despite knowing, or having the opportunity to know, far more. We would feel powerful, with the feeling of knowledge at our fingertips, awaiting our merest whim to become present as understanding in our minds. We would increasingly delegate the consumption of news, at vast scale, to trusted agents who would act on our behalf, who would advise us and who would produce experiences for us. We would feel comfortable and confident in that delegation, and fully in control of it. We would feel as if we – our lives, our personal and business affairs, our potential as human beings – were being managed well, or even optimised, making full use of every fact and every development to improve our situation in the world.
It would feel simple: There would be less ‘stuff’ – less content, fewer artifacts like articles or videos, fewer outlets or sources to check, less tabs or bookmarks to keep track of. We would probably spend considerably less time consuming information, even as we consumed far more information. We would never feel the need to integrate, contextualize and evaluate information ourselves – to effectively act as our own editors in order to make sense of information, as we often must do now. We would not feel the need to hoard information and be far less likely to be constantly engaged in information triage. More context, more guidance, and more information would be delivered to us ‘just-in-time’ – at the moment of need – and we would be better able to direct most of our focus on understanding and on the joining of dots.
It would feel safe: We would, ideally, also feel confident and secure in our information. We would trust our knowledge of events, of facts and our ‘world model’ of society to be accurate and be confident that it represented reality as closely as practicable. We would feel safe, assured that our knowledge was complete, without gaps or blind spots, and without the influence of misrepresentation, fraud, error or incompetence. There would be very few stories, subjects, issues or details that were not ‘covered’ or accessible to us, and so we would feel free to ask any question or follow any curiosity. As we came to expect a position of understanding, rather than a position of ignorance or confusion, we might begin to feel more at home in our world, even during a time of unprecedented change.
It would feel impartial and real: News would feel as if there were nothing and nobody between you and the factual sources of information. There would be no gatekeeping, no agenda setting, no control of narratives and nobody deciding for you what was or was not ‘newsworthy’. In this environment you would not feel the influence of others, because the only influences would be the truth and your own choices. You would never feel spoon-fed or patronised. You would never find yourself parsing through a carefully spun, highly managed narrative, trying to read between lines to understand what was really going on. You would never wonder about the stories that that someone else decided not to tell, because every conceivable story would be accessible to you. You would never be forced to blindly rely on credentials, affiliations, titles or other certifications as a substitute for trust and track record. This independent, direct and transparent access to information on your own terms would feel fair and impartial. You would become far more inclined to ask ‘am I wrong?’, less likely to seek a self-serving story that fudged the facts and more likely to have an expectation and desire for an accurate description of reality. You would feel that you could handle the truth, and we would perhaps all feel like more honest people in a more honest society.
It would feel natural and deeply satisfying: This ideal news environment would satisfy our deep cognitive needs for information, needs that originated in our evolutionary environment over hundreds of thousands of years. Our informational itches would be scratched, our subconscious needs salved. Those who value building and maintaining an internal worldview or world model would feel their understanding grow. Those who value news as a social object around which they can engage with others would know what to talk about with whom. Those who value news for its utility would know exactly what to do. Far fewer people would feel ‘stupid’ or uninformed or excluded because of lack of information or understanding. Far more would feel knowledgeable, and also confident in that knowledge. This relationship with information might even, over time, come to resemble the relationships our ancestors had with information. Their situation was defined by a scarcity of information and by a relatively unchanging society, enabling them to feel mastery over the little information available. Our situation is different, defined by an abundance of information and by a society changing at an accelerating pace, however we might be able to use our AI tools to eventually approach that same feeling of mastery. This might, as we transitioned into this way of knowing, enable us to lead lives with a more stable informational foundation – lives that felt more ‘fundamental’ or ‘intentional’, because they would be built on a richer and deeper awareness of the environment in which they were lived.
It would feel timely, lively and current: We would feel that we were continually up to date, on stories, on topics, on culture and on communities and their conversations. We would never feel that we were missing the latest information about any story or subject or conversation, because focused information would come to us nearly instantaneously if we wished it. We would lose the pressure to ‘always keep up’ because we could effortlessly ‘catch up’ at any time using personal tutorials, summaries and guidance covering exactly what we had missed. We would have the informational currency to act more quickly, becoming more decisive as we became more confident of context and options. We would lose the idea of news as a constant stream, at constant volume, and become far more in touch with the ebb and flow of events – focusing more when more is happening and less when less is happening. All news could be ‘slow news’, or ‘fast news’, as we choose or as our personal situation required.
It would feel shared and inclusive: The information that we consumed would primarily in the context of other people. Common narratives would form and evolve, but social sharing of information might become more focused on common facts, common trust, and a common ability to contextualize stories within the interpretations and arguments of others. We would know alternatives and have a far greater awareness of the difference between objective truth and ‘inter-subjective’ (or commonly assumed) truth. We would know why other people believe what they believe and feel more inclusive towards them. We would know more about what really interested others and thereby improve the quality and value of what we shared then and of what others shared with us. We would know the cultural and informational ‘weather’ of our communities or society, with a finger on the pulse of social media, of information flows and of social sentiment. People, in this ideal news ecosystem, would know where they belonged and would therefore feel that they belonged – ‘finding your tribe’ would become easy. They would know their allies, opponents, assets, liabilities and their options and constraints as well as how they could best optimise their lives towards their goals and values. We would also feel that others in our society were equally well informed, and as a result we might be more respectful of others and their opinions. Similarly, we would feel that any information or point of view that we chose to share with the world would be likely to make its way to others who might find it useful, and therefore we would feel as if our voice in society were fully valued.
It would feel intimate: There would be nothing that you could not know about, no matter how unpleasant, uncomfortable, embarrassing or socially unacceptable. The deep personalisation of every story could extend to the context of your most personal needs and wants, in the same way that a conversation with a therapist, or a doctor or with someone that you trust completely does. You would have answers to your awkward questions, or your selfish questions, or those questions you felt too embarrassed to ask. You could ask for brutal honesty, if that was what you wanted, and that would be provided it to you too. This intimacy would also apply in the other direction, as your informational agents gradually understood more and more about you and your most private information needs and wants. Over time your information consumption might lose much of the distinction between ‘public’ and ‘private’ issues, with both becoming intertwined in ways that made information consumption, even of politics and world affairs, feel integrated with your essence.
It would feel like it mattered: You would only ever see news that felt viscerally useful and relevant. You would never feel as if you were sorting through a stream of uninteresting content looking for a few ‘nuggets’ of relevant information. Instead all information would contribute something new and valuable to you. Every fact and update would build on your understanding or grow your power, and every take and insight would be genuine, meaningful and fresh. You would feel that there was a high bar for news to reach you, and if that bar were breached with irrelevant or useless news then you would quickly correct that. The personal return that you received on your investment in news, and in the consumption of news, would dramatically increase. Each unit of information would make a difference in your life in some way, and you would feel that you would miss it if it were not there. News would feel practical, providing a prescription and guidance for action. Over time it would feel more like wisdom, or at least more like the raw materials for wisdom.
A North Star for News
Every field of human endeavour has an ideal. For medicine it’s a world without cancer, diabetes or Alzheimer’s. For energy it’s a world of limitless clean electricity, available to everyone at marginal cost. For space it’s cities in low earth orbit, and then the moon, Mars and the stars. For AI its artificial general intelligence and then super-intelligence. For too long the ideal for journalism has been the past – that glorious era of gatekeeping power and the status and monopolistic economic rents that went with it. That ideal is no longer credible. We are standing on the cusp of unprecedented transformation in our ability to work with information, driven by AI, and we therefore need a clear vision of what it is we want societal information flows to become. We need an ideal that looks unapologetically forwards towards a meaningfully better future and that brings with it the ambition to work towards its realisation.
My own proposal for this vision, described in this essay, is undoubtedly seriously flawed in many ways. The uncertainty and complexity of our present moment also means that likely none of us are currently anticipating or imagining the way in which our AI-mediated information ecosystem will ultimately play out. I will, however, vigorously defend one specific aspect of my vision for news – its plausibility. Given the AI models and agentic capabilities that we already have and given the capabilities that we are likely to have as a result of the enormous investment currently pouring into AI, we already have technical ‘line-of-sight’ to the experiences I’ve described. Yes, there are enormous challenges to overcome in building this vision, especially around verification and explainability at scale as well as around interface design, intentional and unintentional bias, privacy and many other issues. And yes, we still have many unanswered questions about how audiences – people and citizens – will ultimately choose to use AI and to communicate with each other using AI. These problems, however, are merely work to be done – a natural and normal part of building something better. A greater challenge is to envision clearly and specifically what it is we might build, and how our lives and societies might become better because of it.
We urgently need a real societal conversation about what news and journalistic information should ideally look like in an AI-mediated world – a conversation that includes voices beyond established journalists and the legacy news industry. Technologists, engineers and scientists have significant contributions to make here, as do investors, venture capitalists, economists, philosophers, designers and creative thinkers at the avant-garde of policy, business and culture. We especially need voices from the information sciences and cognitive sciences. It would be good to also hear from younger and less privileged people, who are often ignored but who have the greatest potential to contribute fresh insight.
If you are engaged in any way with finding an AI future for societal information, then you should join this conversation. You should be thinking deeply about your own ideal experience of news and information, about what you would like to know about the world and about how you would like that knowledge to make you feel. You might also consider sharing your vision with the world, adding your voice to thoughtful contributions from people like Juan Carlos Lopez Calver of Schibsted News Meda, Shuwei Fang of Harvard’s Shorenstein Center, or Joshua Rothman in The New Yorker.
As I mentioned at the beginning of this essay, this article is the first in a series. In the next piece I will look at the prospect of societal information as a dynamic, fast-growing sector of the global economy, attracting energy, investment and talent, and motivated by the potential of a radically more functional society that is informed and empowered by a fundamentally transformed information ecosystem. That prospect will build on the vision described here and examine how the global innovation community – including Silicon Valley – might approach societal information de novo, instead of burdened by the assumptions and interests of legacy media. It will attempt to envision an approach and perspective for innovation in news that imagines, invests and builds in the same way that innovation in medicine, energy, space or AI itself does.
From our vantage point in 2025 we can look back over centuries of innovation in how society informs itself and appreciate how much richer our live are because of it. We now need to look forward. AI offers us a truly miraculous toolkit that may exceed the internet or even the printing press in its potential. We need to decide what we want to do with it.
Wow, oy and ... wow. That's an exhausting... er, I mean exhaustive list of everything we COULD know in this brave new AI world. But WOULD know? Just how many brain-cell augments (and some new sense of time itself) would it take to make that in the least bit feasible? Or do we use the AI robots in our lives as the ultimate curator, and keep track of alllllllllllll that stuff so we can learn of it if, how and when we wish or need to? The mind boggles, as it does so often in today's AI-related discussions. Thanks!
Hi David, this recent piece by The Guardian’s Katherine Viner resonated with yours. She argues that the way out of AI-mediated and flooded digital ecosystem is a human-centric approach. That does fail to address the political economy of the industries themselves - including journalism (while the Guardian is an exception) . Thought you might enjoy it. Looking forward to our discussions at the summit https://www.theguardian.com/media/ng-interactive/2026/may/06/how-to-survive-the-information-crisis-we-once-talked-about-fake-news-now-reality-itself-feels-fake