How Will AI Change Local News in the Next 5 Years?
By Adriana Stephan and Claire Leibowicz
Despite the vital functions local newsrooms serve, such as covering hyper-local issues unreported by national publications and directing their audiences to vital resources in their communities, the local news industry is in dire straits. Since 2005, around 2,200 local newspapers across America have reportedly closed amid declining ad sales and subscription rates. As these newsrooms are forced to do more with fewer resources, some have looked to AI and other emerging technologies as potential solutions that could augment reporting and improve the bottom line of news businesses. But what impact will AI tools have on the local news landscape, and how can they be used both ethically and responsibly?
To begin answering these questions, the Partnership on AI (PAI) asked nine experts across media, academia, civil society, and industry, “What is the most radical change AI will bring to how local news is produced, distributed, and/or shared online by 2026?” While the responses that follow will be of particular interest to journalists and technologists, the importance of local news to public understanding makes these answers worth considering for everyone with a stake in the health of local news and our information ecosystem more broadly.
Preserving a healthy information ecosystem requires reconciling difficult tradeoffs, including how to maximize the potential of using AI to support local journalism while minimizing the risks. For example, AI can provide more personal insights about your community, but it can also produce information silos and unrepresentative newsfeeds.
This blog post represents the first output of PAI’s recently launched AI and Local News Workstream supported by the Knight Foundation. Building on previous work by PAI’s AI and Media Integrity Program on bolstering the quality of public discourse, this Workstream is developing ethical best standards and practices for AI’s use in local news. Doing so in a way that will sustain local news in the long term requires collecting input from a diverse array of stakeholders representing multiple perspectives, work that this blog post will discuss in more detail at the end.
What is the most radical change AI will bring to how local news is produced, distributed, and/or shared online by 2026?
Understanding the speed at which technological change can occur is sometimes only possible in a rear-view mirror. In 1993, I still connected into the Internet using a dial-up modem and online newspaper products were rare. In 1995, came Craigslist. By the year 2000, we had the Web, Yahoo and Google. By 2005, we had Facebook and YouTube. Along the way, journalists began to work in increasingly sophisticated ways, especially in regards to data and computational journalism. Now, I believe we are approaching a level of knowledge and integrated knowledge-sharing across domains that will foster rapid change.
So, by 2026, what can we expect? Already, large news organizations are using automated methods to produce stories or news alerts, for example, the L.A. Times and its earthquake alerts, or earnings reports by the Associated Press and Reuters. But AI, in particular the use of machine-learning to mine text for meaning, is about to take off. This leaves me sanguine about the possibility of supporting local news. Big Local News is building Agenda Watch to mine local government agendas and minutes, and partnering with local news organizations, for just this purpose. But I believe the possibilities in this area will continue to expand in ways we haven’t even conceived of yet. I can envision using AI for better ingestion of information into structured data, such as with police misconduct records. Or, for organizations like Outlier Media to better inform very local users of decisions or events that will impact their lives, for example, utility rate hikes, or policies for rent relief.
Many problems in AI are still hard, but even those — mining key information from video for example — are getting easier. Take a look at the Cable TV News Analyzer by Maneesh Agrawala of The Brown Institute for Media Innovation as another example.
Where I see the most potential for advancement though is not in the world of pure AI, but in human-directed AI, where machine learning allows for filtering of large amounts of documents or data. Human users can then review those results, use semi-automated methods to rinse, repeat, and then extract journalistic meaning out of the work. The bottom line will be increasing possibilities for humans to use a suite of powerful tools to help support very local accountability reporting and patterns that are regional, national or even international in scope.
But as noted by my Big Local News colleague Serdar Tumgoren, there are challenges lurking around the corner too.“I often worry that AI will also pose major challenges to the news and information ecosystem in the near future. Most notably the ability of state actors and others to generate and distribute disinformation on a massive scale using algorithms such as GPT-3,” Tumgoren told me. “The news and more generally the public will likely need to turn to AI to help combat the very challenges posed by AI.”
Founder and Director of Big Local News, Director of the Stanford Computational Policy Lab, Hearst Professional in Residence, Stanford University
There are many places to turn for local news, from legacy newspapers and TV stations to neighborhood Facebook groups. If a busy person doesn’t have a news source they trust, it can be daunting to sift through those options.
In the next five years, tech leaders and journalists should work together to ensure high-quality local news shows up when people are searching for information about their community.
Local newsfeeds powered by AI do not currently provide an accurate representation of many places. Stories about crime are over-represented. Stories about what’s going well and how community members are responding to social challenges are missing.
Journalists and the tech industry share responsibility for this problem and a human-centered solution is required for both.
Newsfeeds on platforms like Google and Facebook that use a set of keywords or zip codes may not fully represent how people define their local community. For example, communities that are far apart, like El Paso, Texas and Las Cruces, New Mexico, can actually share deep cultural and economic connections that make people interested in news from both places. At the same time, many news outlets have not covered all groups equally, leaving a gap in the kinds of stories people are seeking.
One approach that could dramatically shift our thinking about the definitions of local news would be to ask people to help us map how they see their local community, along with its assets and challenges. Those insights would be beneficial for both industries.
We may live in a global world, but people will always be curious about what’s happening around the corner, in their local schools and local government. AI can do more to elevate trusted local news organizations and ensure people have the information they need.
Sarah Gustavus Lim
New Mexico Local News Fund Founder and Economic Mobility Manager at Solutions Journalism Network
Over the next 5 years, local news faces the same product objective they’ve faced for centuries: how to help people better enrich and engage their lives within their communities. AI can further accelerate this work, using intelligent algorithms to help provide readers with increasingly smarter and relevant insights into the neighborhoods they live in. Learning about the people and issues in your area — or even finding out about a new restaurant — is an important part of building community.
In 2026, this could mean a person wakes up in Islington, London, and is greeted with a tailored news summary specific to their day and their interests. First, an algorithmically assembled alert that their bus line is likely to be delayed several minutes due to roadwork (created from a merger of London transport data, paired with their geo-location data). Next, an advertisement from a local venue that a band they’d like to see is playing tonight, and tickets are 10% off. A reader-reported insight comes up next: There was a car break-in 0.2 miles from their apartment yesterday, but the data widget beside the insight puts in context that car break-ins are actually dropping in their specific neighborhood. Finally, they see an in-depth article about air quality in London, a topic they read about often, by their favorite journalist; this is next to an interview with their recently elected Islington council member.
The above is just one imagining of how AI could help readers better navigate their day; as we push towards this future, there are many open questions to resolve. For instance, how do we protect against information silos and give users the larger view of the health and well-being of their societies? And as news integrates more structured data (crime etc.) how can we ensure the accuracy of that data and give it appropriate context? How do we counter biases in AI models trained on insufficient data (e.g. a dataset’s lack of diversity in race or gender)?
Which portion of this work is handled on the platform side vs. publisher side remains uncertain as well. Local publishers have the local knowledge to identify and put into context relevant sources of local data, the relationships and proximity to work closely with local advertisers, and the reporting expertise to determine which stories require a deeper editorial investigation. Platforms have expertise on how to best distribute the right content to the right reader, as well as deep knowledge and investments in AI. The result will hopefully be a collaboration between the news industry and platforms to create a better user experience for readers, closer relationships between users and publishers, and better informed local communities.
Product Manager, Local News Experiments, Google
There are some amazing developments using AI in local news. By 2026, AI will better transcribe and translate local news videos for diverse audiences around the world. It will also automatically tag local content to be better searchable by algorithms, which may lead to the higher online visibility of local news. AI-powered tools could help newsrooms achieve their ethical goals by, for example, monitoring whether journalists are quoting from equal numbers of men versus women. For resource-strapped community newsrooms, these tools can provide creative ways to meet audience needs. They may even translate to more revenue at the local level.
What won’t change is the deeper ethical and political questions that AI presents. AI algorithms make decisions, but who controls AI? As newsrooms take pictures of crowds at political demonstrations or rallies, what prevents governments or even non-state actors from benefiting from the large-scale image recognition used to determine who attended? Or, in making local news more visible, what does it mean for newsrooms to make editorial choices when social media algorithms control how a topic trends?
The appropriate role of AI-related technologies, including news, upon our own individual and collective decision-making is far from clear. AI may improve local news quality, but the systemic tendency of internet distribution strikes me as overall controlling, or authoritarian. It takes a really, really “tough mind” — Martin Luther King, Jr.’s phrase amid an equally confounding context — to successfully navigate the information riptides of our current online environment. We need to keep on asking: How can improved AI create news that aids local citizens with their everyday decision-making? Because, right now and for the foreseeable future, many choices in how content is shared, viewed, and even created will not be made by citizens or by newsrooms but by algorithms.
Connie Moon Sehat
Media consumers in many parts of the world, especially those coming online for the first time, are often exposed to an information diet devoid of the foundational understanding of media and information literacy that content produced and shared through mainstream media ecosystems — print, radio, television — provides. In the geographies where closed, encrypted networks are first ecosystem media technologies, a front that is largely inaccessible to fact-checkers, AI can help with just-in-time context and media and information literacy.
As such, we must advance policies, technologies, and corporate commitments to developing systems and algorithms that preserve privacy while creating a path for end-users and their communities to seek and share contextual information. The role of AI in serving content and context to community moderators and concerned end-users is going to be a critical aspect of addressing information integrity in closed messaging spaces. Local news providers will increasingly be citizen journalists whose editorial decisions are augmented by algorithmic analysis of questions and concerns moving through messaging networks.
AI can guide these decisions, providing a way to “bake in” contextual information, at the point of information generation. Local news is likely to benefit greatly from this, using AI to identify local patterns and trends of interest for further examination. This system may not work in all contexts, and in news deserts, it could potentially amplify misinformation and knowledge gaps, leaving room for potential manipulation. It is therefore important to recognize any potential gaps and to address them going forward. In this way, AI can therefore contribute meaningfully to robust media ecosystems, putting credibility at the heart of information generation and adding meaningful layers to information that would otherwise be lacking.
The same challenges we address in closed messaging spaces do, we believe, port to new local media consumption patterns emerging in metaverses, autonomous vehicles, and anywhere computation is augmenting our lived experiences. Empowering users to access, assess, and share context in these new local media environments should be a critical concern of policy-makers, designers, and technologists alike.
Founder and CEO, Meedan
Program Manager, Africa, Meedan
The question I ask is a different one: What is the most radical change data-based technology can bring to the local news enterprise? AI is one type of data-powered technology. Because it can generate predictions, it gets hyped as an all-powerful inevitable force. But the real challenges for local news organizations are in the socio-technical realm.
Local newsrooms have to reframe their whole journalistic relationship to and with the multicultural communities who form not just their “audiences” but also participants in journalism as a civic process. This will allow local newsrooms to defend their journalism as methodologically robust (accuracy), democratically grounded, reliable and relevant. Several American newsrooms are leading the way in bringing in new norms for their journalism: engagement, solidarity, and inclusion. It is when strategizing around this transformative journey that clarity on the role for data and data-based tooling systems like AI and ML can help.
Take this question: “Are stories that score much higher on DEI-quality and representation actually helping local newsrooms diversify audiences systematically and predictably?” As a way to drive strategy, this question is not easy to answer on a management dashboard today. That requires zooming into three feedback loops already running in many local newsrooms.
1. Journalists <-> Community: All the interactions between people, reporters, and their stories, online or offline, including story engagement, comments, social media play, leads, etc.
2. Community <-> Business: Where the enterprise is connecting with audiences to raise money or get people to subscribe, pay for services, particular premium offerings, etc.
3. Business <-> Journalists: The main backend feedback loop with the news organization on how stories are performing. For example, is a particular story series leading to improving some business outcomes like new subscriptions or a new advertiser/brand sponsorship?
To answer this type of question and more like this, local newsrooms need to connect these three feedback loops between business, journalism, and community using ethical data-tooling and translation. AI is good at helping understand connections within diverse datasets as a means to an end, but always risks flattening meaning when used blindly.
I’d like to see a future where local newsrooms uniquely define new data that each feedback loop can take in from the others, gather, translate and label this ethically, and share into a common strategic arc for the whole organization. Incentivizing ethical local journalism long-term requires data about stories, their sourcing, their comments/engagement, and business metrics to all be interconnected in a single chain. Principles and guidelines from the data and AI ethics communities can provide guidance so that the interconnected analysis system itself does not become extractive and unethical. This is a serious and pro-democracy opportunity waiting to be unlocked here for enterprise local news.
Subbu Vincent, @subbuvincent
Director of Journalism and Media Ethics, Markkula Center for Applied Ethics, Santa Clara University
In 2004, Facebook was new, Netflix a DVD business, the Apple iPhone didn’t exist, and I graduated from journalism school. Going into local TV news seemed like a solid career choice. I couldn’t anticipate the rapid shift in how people consumed video news because we didn’t have broadband connections. I only had a DSL line at home.
How quickly things changed. Using what I’ve learned over the last 15 years in local TV news and armed with the knowledge from a fresh master’s degree in business information systems, I feel I can confidently predict… what I will have for lunch today. Take my 2026 predictions with a degree of skepticism.
The socio-economic divide reflected in broadband speeds will continue to impact how people consume local TV news. In 2026, TV stations in markets with easy access to high-speed internet will lose more viewers to digital substitutes. AI can help personalize local news for digital audiences and expand the scope of coverage with more automated content. In a digital world, TV newsrooms are free from the constant pressure of needing video for every story. This puts broadcasters on a catch-up footing with our print cousins, who I guess have a 10-year head start in digital maturity.
In 2026, TV stations in markets with slow internet speeds won’t see much of a change from today. The local broadcaster will remain a vital source of quality news throughout all dayparts. There will be little incentive to invest in digital without the infrastructure and audience. AI can still help journalists in newsgathering and production. Everyday needs include transcription and shot sheets. Freeing journalists from those low-level tasks lets them focus on putting in extra time on writing and producing that wins awards.
AI Product Manager, Associated Press
Local newspapers have clearly been decimated over the last couple decades. At the same time local news sites, blogs, and even platforms have also popped up — in my home of Chicago, the online site Block Club Chicago is a worthy source alongside the stalwart Chicago Tribune. Yet it’s hard to shake the idea of news deserts, a metaphor which is both salient and alarming. I imagine thirsty souls dragging themselves through barren terrain seeking a drop of information to sustain themselves and their communities. I’m lucky to live in a city with local information options — smaller cities and rural areas are surely at a disadvantage. But are there really local news deserts? AI can help answer this question.
Scholars have known for some time that we need methods and measures to assess whether people’s critical local information needs are being met. This includes things like knowing about public safety, health, education, transportation, environment, economic development, as well as civic and political life. While there are already several initiatives, from GDELT to MediaCloud or even the Common Crawl, that gather and index news media, AI can be layered on top to measure where and to what extent there is adequate supply of local media to meet these critical needs. Together with suitable computational definitions of local news, classifiers can be tuned to pick up on the true supply of local information, regardless of whether it’s from a local newspaper.
By measuring local news supply at scale, AI will give us the tools we need to manage the provision of local information for our communities. Measures of supply, tracked over time, and broken out by critical information needs met, will offer local media makers, advocates, and policymakers a map that exposes the true gaps in people’s local information ecosystems.
Associate Professor in Communication Studies and Computer Science (by courtesy) and Director of the Computational Journalism Lab at Northwestern University
Over the next two years, the AI and Media Integrity Program will continue to convene a multistakeholder community to address the questions and concerns raised above and develop recommendations to ensure local newsrooms and platforms are embedding responsible AI in their news production and distribution practices.
As these experts note, AI will help journalists mine text for meaning. It also risks over-concentrating decision-making power in the hands of algorithms. These responses provide an overview of the promise of emerging technologies in local newsrooms, highlight the ethical challenges of its implementation, and illustrate how we might drive toward ethical norms if we proactively engage with ethical questions now.