The Future of AI and Journalists
Reuters used AI to expose a mass grave in Syria. The Associated Press has been using AI to write earnings reports since 2014. And in 2025, 82% of news executives said their organisations already use AI for news gathering, with 97% calling end-to-end automation essential. At the same time, newsrooms have lost a third of their search traffic in a year, AI-generated disinformation is scaling faster than any fact-checking operation can keep up with, and 67% of publishers say AI has not saved them a single job yet despite massive investment. The story of AI and journalism is not a simple one about robots replacing reporters. It is a story about an entire industry being restructured — some of it for the better, some of it at genuine risk — while the public's need for reliable information has never been greater.
Table of Contents
- What AI Is Already Doing in Newsrooms
- The Traffic Crisis Nobody Is Talking About Loudly Enough
- The Journalism Jobs Picture
- What AI Cannot Do in Journalism
- The Misinformation Problem
- How AI Is Actually Helping Investigative Journalism
- What Journalists Should Do Now
- What the News Looks Like in 2030
- Frequently Asked Questions
What AI Is Already Doing in Newsrooms
The transformation of newsrooms by AI is already well underway — and further along than most coverage of the topic suggests. The question in 2026 is not whether newsrooms are using AI, but what they are using it for and whether it is working.
The 2026 newsroom reality: 82% of news executives say their organisations already use AI for news gathering. 97% consider back-end automation essential. 87% say their newsroom is being "fully or somewhat transformed" by generative AI. 75% of executives expect agentic AI tools — autonomous systems capable of multi-step tasks — to have a large or very large impact on newsroom operations in 2026. Yet just 44% say their AI initiatives are showing promising results. 42% describe the impact as "limited so far." Source: Reuters Institute for the Study of Journalism, Trends and Predictions 2026.
What AP and Reuters figured out early
The Associated Press began using AI to write corporate earnings reports in 2014 — a full decade before most newsrooms were discussing AI seriously. Automated Insights' Wordsmith platform produced thousands of standard earnings summaries per quarter, freeing AP business reporters to cover the stories behind the numbers rather than the numbers themselves. Reuters has gone further: its News Tracer tool monitors social media in real time to identify breaking news stories before they become widely known, and its Lynx Insight platform analyses large datasets to suggest angles and story ideas. Most significantly, Reuters built custom AI tools to process tens of thousands of documents from Syria's security forces — translating, indexing, and searching them to expose evidence of atrocities that no human team could have processed at that scale.
What most newsrooms are doing with AI in 2026
The most common applications are back-end and workflow — not the front-page, bylined stories that most readers see. AI writes headlines and summaries, generates newsletters, copy-edits drafts, transcribes audio, translates content into multiple languages, monitors breaking news feeds, and produces the standard-format stories — weather reports, sports scores, financial results, traffic updates — that follow predictable patterns. These are genuinely useful applications. They save time on tasks that are important but not what most journalists entered the profession to do.
The agentic AI shift of 2025–2026: From 2023–2024, newsrooms automated individual tasks. From 2025, the shift is toward agentic AI — autonomous systems that handle end-to-end complex workflows without step-by-step human instruction. An agentic AI system does not just write a summary; it monitors a topic, identifies a developing story, gathers relevant sources, drafts the story, checks it against the house style guide, and queues it for editor review — autonomously. This is a qualitatively different kind of automation from what came before, and 97% of news executives consider it essential for the near future.
The Traffic Crisis Nobody Is Talking About Loudly Enough
The most significant impact of AI on journalism in 2026 is not the stories AI writes. It is the readers AI is taking away.
Google's AI Overviews — which appear in approximately 10% of US searches — have accelerated a shift that was already underway: readers getting answers from AI summaries without ever clicking through to the news article. Chartbeat analytics covering more than 2,500 news sites show Google organic search traffic down 33% globally and 38% in the US between November 2024 and 2025. Publishers expect search traffic to decline a further 43% over the next three years. One in five expect a decline of more than 75%.
What this means in practice: The content most at risk from AI-driven traffic loss is the content that has funded newsrooms for decades — lifestyle, service journalism, weather, TV listings, evergreen explainers, and local news. These formats are being absorbed into AI answer engines that answer the query without directing anyone to the original source. What is more resilient is investigative reporting, original analysis, exclusive interviews, and content that requires human judgment, sources, and accountability. The distinction between commodity content and distinctive journalism has never mattered more — or been more commercially urgent.
The Journalism Jobs Picture
The employment picture in journalism is complicated, and the AI contribution to it is more nuanced than a simple displacement story.
Where journalism employment is holding or growing
- Editorial-tech hybrid roles — Journalists who can work with AI tools, understand data, build workflows, and bridge editorial and technical teams are in demand. This is one of the fastest-growing role categories in major newsrooms.
- Investigative journalism — AI is actually expanding the scope of what investigative teams can do, making large-scale data journalism more accessible and enabling investigations that would previously have been impossible.
- Verification and fact-checking — As AI-generated content scales, demand for skilled verification journalism is growing. Organisations like Full Fact and IFCN-affiliated fact-checkers are hiring, not cutting.
- Audience and product journalism — Roles focused on direct reader relationships, newsletters, community engagement, and subscription growth are expanding as publishers shift away from traffic-dependent models.
Where journalism employment is under pressure
- Commodity content production — Writers producing standard-format articles following predictable templates face direct competition from AI that produces the same content faster and at lower cost.
- Wire service and aggregation roles — Jobs built around monitoring, summarising, and redistributing content from other sources face the most immediate automation pressure.
- Local news — Already in structural decline before AI, local news organisations face the most acute version of the traffic and revenue crisis — with less capacity to invest in the distinctive journalism that provides protection.
- SEO and content marketing journalism — Roles producing high-volume web content primarily for search engine visibility face severe pressure as AI produces equivalent content at near-zero marginal cost.
The honest employment picture: 67% of publishers report AI has not saved them any jobs to date, and only 16% report modest staff reductions from AI. This is not because AI has not changed what people do — it has. It is because most newsrooms have absorbed AI's productivity gains without reducing headcount, using the efficiency to produce more content, cover more topics, or improve quality rather than to cut costs. Whether this continues as AI capability grows is the central employment question for journalism over the next five years.
What AI Cannot Do in Journalism
The clearest framework for understanding AI's role in journalism is the distinction between commodity content and distinctive journalism. AI is very good at the former and cannot, in any meaningful sense, do the latter.
- Build sources and relationships — The most important journalism — the stories that hold power to account, reveal wrongdoing, and change things — depends on sources who trust a specific journalist enough to share information at personal risk. That trust is built over years of human interaction and cannot be replicated by an AI system.
- Bear legal and ethical accountability — A journalist who publishes a defamatory story can be sued. An editor who approves a fabricated story faces professional consequences. An organisation that publishes without adequate verification faces reputational damage. These accountability structures are what give journalism its authority. AI has no professional accountability — it is a tool. The human journalists and editors who use it retain all the accountability.
- Exercise genuine editorial judgment — What makes a story important? Which angle serves the public interest? When is a source credible? When does a story need more reporting before it is ready to publish? These judgments require contextual understanding, ethical reasoning, and a developed sense of what matters — none of which AI currently provides.
- Conduct original investigations — Investigative journalism requires identifying that a story exists in the first place, building the source relationships to confirm it, navigating legal risk, making editorial judgment calls about what to publish, and taking accountability for the consequences. AI can assist with parts of the process — data analysis, document processing — but cannot drive it.
- Earn audience trust — In an information environment flooded with AI-generated content, the thing that distinguishes journalism audiences trust is precisely its human origin and human accountability. Audiences continue to show a strong preference for journalism authored by humans, despite AI's ubiquity in newsrooms. This preference is not irrational — it reflects an accurate understanding of where accountability lies.
The Misinformation Problem
AI's most significant negative impact on journalism is not the jobs it may eliminate. It is the disinformation it is already generating at scale.
The scale is growing fast: In Brazil, the fact-checking organisation Aos Fatos found that 16% of the 619 claims it checked in 2025 involved AI-generated content, up from 7% the previous year. AI-generated fast content reached over 32 million views across TikTok in Brazil alone, with 2.1 million interactions on Facebook and Instagram related to AI-powered disinformation. The growth is driven primarily by fabricated visuals — AI-generated images and video increasingly indistinguishable from real footage. Full Fact CEO Chris Morris put it plainly: "We are in danger of getting to a place where no one believes anything they'd read or see or hear anywhere."
The challenge for journalism is that AI-generated disinformation scales faster than any human fact-checking operation can respond to. The same technology being used to produce false content is now being used to fight it — AI-powered fact-checking and verification tools can process far more claims than human fact-checkers. But the scale asymmetry remains: producing disinformation with AI is far cheaper and faster than verifying it, even with AI assistance.
| Content type | AI threat level | Journalism's protection |
|---|---|---|
| Breaking news from official sources | Moderate — AI can fabricate quotes and events | Source verification, official confirmation |
| Photographic and video evidence | High — AI-generated visuals increasingly convincing | Metadata analysis, provenance verification tools |
| Statistical and data claims | Moderate — AI can generate plausible-looking false data | Primary source checking, data journalism skills |
| Investigative findings | Low — genuine investigations are hard to fake | Accountable sourcing, document verification |
| Opinion and analysis | High — AI produces fluent, confident-sounding analysis | Named authors, track records, editorial standards |
How AI Is Actually Helping Investigative Journalism
The most underreported story about AI and journalism is not how AI threatens reporting — it is how AI is enabling journalism that would previously have been impossible.
The Reuters Syria investigation is the clearest example. When reporters on the ground obtained tens of thousands of documents from Bashar al-Assad's security forces, the sheer volume made human processing impractical. Reuters journalist Allison Martell built custom AI tools to translate, index, and search the documents — enabling the team to expose the regime's plan to move a mass grave to hide evidence of atrocities. This is the kind of accountability journalism that AI's critics are worried it will replace. In this case, AI is what made it possible.
The same pattern appears across investigative journalism globally. A Nigerian newsroom used AI tools to analyse flooding data for a major investigation. Norwegian journalists used AI-assisted analysis of Ukrainian war documentation. The Pulitzer Center has been supporting data journalism projects where AI analysis of large datasets enables stories that would have taken years to produce manually. Tools like Bellingcat's open-source intelligence methods — combining AI analysis with human verification — have become a model for how newsrooms can investigate complex global stories at scale.
The key distinction: AI is a powerful amplifier of journalism that requires human judgment, sources, and accountability to initiate and direct. It is a weak substitute for journalism that requires any of those things to produce. The newsrooms that are using AI most effectively are using it to expand what their journalists can investigate — not to replace the journalists doing the investigating.
What Journalists Should Do Now
- Develop genuine AI fluency — Not surface-level familiarity, but the ability to use AI tools effectively for research, data analysis, document processing, and workflow automation. Journalists who can use AI to do things that would previously have required a data team are significantly more valuable than those who cannot. The Reuters Institute identifies editorial-tech hybrid as the fastest-growing role type in major newsrooms.
- Invest in the things AI cannot do — Source relationships, subject-matter expertise developed over years, the track record that makes sources trust you with sensitive information, the legal and ethical judgment that distinguishes credible journalism. These are the career assets most resilient to automation and most valuable in an environment flooded with AI-generated content.
- Become a specialist in something real — Generalist news writing is the category most vulnerable to AI automation. Deep expertise in a beat — science, finance, politics, health, technology — where a journalist builds contacts, understands context, and can evaluate claims that non-specialists cannot, is substantially more durable.
- Understand verification as a core skill — In an information environment where AI-generated content scales faster than any human can read, the ability to quickly and rigorously verify claims, images, and documents is becoming one of the most valuable skills in journalism. Organisations like the Poynter Institute offer specific verification and fact-checking training that is growing in demand.
- Build direct audience relationships — The journalists navigating this transition most successfully are those who have built audiences that follow them personally, not just the publication they write for. Newsletters, podcasts, community engagement, and the cultivation of a personal editorial identity are all ways of building the direct relationship with readers that is less dependent on search traffic.
For broader context on how AI is reshaping professional careers, see our guides on how AI is transforming the legal profession, what jobs AI will replace, and our beginner's guide to AI.
What the News Looks Like in 2030
- Now — 2027 (Structural transition): Search traffic continues declining as AI answer engines take a larger share of information queries. Publishers double down on direct subscription relationships and newsletter audiences as the traffic-dependent model erodes further. Agentic AI handles most commodity content production autonomously. Local news organisations face the sharpest crisis — those without distinctive journalism or direct community relationships face existential pressure. AI-generated disinformation reaches unprecedented scale, driving investment in verification infrastructure. Editorial-tech hybrid roles become the most sought-after in major newsrooms.
- 2027–2030 (New equilibrium taking shape): The distinction between publishers with distinctive journalism and those producing commodity content becomes a survival question, not just a quality one. Subscription and direct revenue models mature. AI-assisted investigative journalism expands — teams of two doing the work that previously required ten. Regulatory frameworks for AI-generated news content and labelling requirements begin taking shape in the EU and some US states. The most trusted news brands are those that have most visibly committed to human accountability and verification standards.
- By 2030 (The settled picture): Journalism has bifurcated more sharply than at any point in its history. One part of the industry — high-volume, commodity, automated content — is largely AI-produced, with humans in oversight and editorial direction roles. The other — original reporting, investigation, analysis, accountability journalism — is more human than ever, more valued than ever, and funded by audiences who have learned to pay for what they cannot get from an AI chatbot. The total number of journalism jobs is lower, but the jobs that remain require more skill, carry more accountability, and are more important for democratic society.
Frequently Asked Questions
Will AI replace journalists?
For some types of journalism, it already has — or is well underway. AI produces earnings reports, sports round-ups, weather summaries, and other standard-format content that follows predictable patterns, largely without human involvement. For the journalism that matters most to democracy — original reporting, investigation, accountability, analysis — AI is a tool that can amplify journalists' capability but cannot replace them. The accountability, source relationships, editorial judgment, and public trust that distinguish journalism from information production are inherently human. The jobs most at risk are those built primarily on producing commodity content quickly. The jobs that are growing are those requiring genuine expertise and judgment.
How is AI already being used in newsrooms?
More extensively than most readers realise. The Associated Press has used AI to write corporate earnings reports since 2014. Reuters uses AI to monitor social media for breaking news and to analyse large document sets for investigations. Most major newsrooms use AI for headline generation, article summarisation, newsletter creation, copy editing, audio transcription, and translation. Back-end automation is considered essential by 97% of news executives surveyed by the Reuters Institute for the Study of Journalism. The shift in 2025–2026 is from task automation to agentic AI — autonomous systems that handle end-to-end workflows without step-by-step human direction.
Is AI-generated news reliable?
It depends entirely on the context and the oversight applied. AI writing earnings summaries from structured financial data, under editorial oversight at a major wire service, is reliable. AI generating news content without primary source verification, editorial oversight, or professional accountability is not. AI-generated disinformation — fabricated news stories, fake quotes, synthesised images presented as real — is growing rapidly and is a genuine threat to the information environment. The presence of human editorial oversight, source verification, and professional accountability is what distinguishes reliable AI-assisted journalism from unreliable AI-generated content.
How is AI affecting news traffic and revenue?
Significantly and negatively for traffic-dependent publishers. Google organic search traffic fell 33% globally and 38% in the US between November 2024 and 2025 across 2,500+ news sites analysed by Chartbeat. Publishers expect further losses of 43% over the next three years as AI answer engines continue absorbing information queries without directing traffic to original sources. The revenue models most at risk are those dependent on high-volume search traffic. Subscription models, direct reader relationships, and distinctive journalism with loyal audiences are proving more resilient than traffic-dependent advertising models.
What journalism jobs are safest from AI?
Investigative journalism, beat reporting requiring deep expert knowledge and source relationships, fact-checking and verification, editorial leadership roles, and editorial-tech hybrid positions bridging journalism and AI capabilities are the most resilient. The common thread is genuine expertise and accountability that cannot be automated. The jobs most at risk are high-volume standard-format content production, wire monitoring and aggregation, SEO-focused content writing, and any role primarily built around producing content that follows predictable templates.
Is AI making disinformation worse?
Yes, measurably. In Brazil, 16% of fact-checked claims in 2025 involved AI-generated content, up from 7% the previous year. AI-generated fabricated visuals are particularly hard to counter because they are increasingly indistinguishable from real footage. The scale asymmetry is the core problem: producing AI-generated disinformation is far cheaper and faster than verifying it, even with AI-powered fact-checking tools. This is one of the strongest arguments for investing in human journalism — the accountability, verification standards, and source relationships that distinguish credible journalism from AI-generated content are precisely what makes it trustworthy.
Should journalism students still pursue the field?
Yes — but with a clear-eyed view of which parts of the field have futures and which are being automated. Investigative journalism, specialised beat reporting, data journalism, verification, and editorial-tech hybrid roles are growing and in demand. High-volume content production for search traffic is being automated. The journalists entering the field now who develop genuine AI fluency alongside strong reporting skills, deep subject expertise, and the source relationships that only come from committed beat journalism will find real opportunities.
What is generative engine optimisation and why does it matter for journalism?
Generative Engine Optimisation — GEO — is the practice of structuring content so that AI systems like ChatGPT, Gemini, and Perplexity cite it as a source when answering user queries. As search traffic declines and more people access information through AI chatbots rather than Google, appearing as a cited source in AI-generated answers becomes a key distribution channel for news organisations. Publishers not thinking about GEO alongside traditional SEO are losing visibility in the channel that is growing as Google search traffic declines. This rewards credible, well-structured, authoritative journalism — one of the more encouraging aspects of the AI transition for quality news.
