AI will not replace journalists, but it is already replacing some journalism tasks. Software can draft routine, data-driven stories, transcribe interviews, and summarize documents, while reporting, verification, source relationships, and editorial accountability still need people. The bigger risk for most newsrooms is economic: AI answers can pull readers and revenue away from publishers.
This guide is for publishers, editors, and newsroom leaders deciding where AI fits. It covers what AI does well, where it fails, what researchers and press freedom advocates are saying, and how to protect the work that only journalists can do.
What is AI in journalism?
AI in journalism means using machine learning and large language models to produce, process, or distribute news. A long-running example is the Associated Press: in 2014 it began automating short corporate earnings reports with Automated Insights, aiming to grow from about 300 manual reports per quarter to as many as 4,400, according to Nextgov.
Today the toolset is much wider. Newsrooms use AI for transcription, translation, tagging, headline testing, summarization, personalization, and first drafts. Most of this work sits around the reporting rather than replacing it.
The replacement question keeps coming up because AI can now produce fluent text in seconds. Readers can ask a chatbot for the news, and that changes how much traffic reaches publisher sites. So the question is really two questions: can AI do a journalist's job, and can newsrooms afford journalists if AI takes their audience?
Will AI replace journalists?
No, but it will change which journalism jobs exist and how many. AI is good at turning structured data into readable text. It is much weaker at deciding what matters, finding out what nobody has published yet, and standing behind a claim.
The research points the same way. Writing for Brookings, Courtney C. Radsch notes that AI cannot go into a courtroom, interview a defendant behind bars, or cultivate a whistleblower's trust. Her conclusion is direct: "the only way journalism will survive AI is to double down on its journalists" (Brookings, 2024).
The United Nations Regional Information Centre (UNRIC) takes a balanced view. Its World Press Freedom Day piece says AI can streamline transcription, keyword searching, trend analysis, and initial drafting, freeing journalists for investigation and fact-checking. It also warns that automation could replace not only reporters but designers, editors, and distribution staff.
What happens when AI writes without enough human checks is already on record. In January 2023, CNET issued corrections on 41 of the 77 stories it had published using an AI tool, The Verge reported. The tool could produce articles, but people still had to catch the mistakes.
What AI can and can't do in a newsroom
The clearest way to answer "will AI replace journalists" is to split the job into tasks.
| AI does well | AI does poorly or not at all |
|---|---|
| Transcribing interviews and meetings | Building trust with sources and whistleblowers |
| Drafting routine data stories (earnings, scores, weather) | Original reporting from courts, councils, and conflict zones |
| Summarizing long documents and translating | Deciding what is newsworthy and why it matters locally |
| Tagging, metadata, and headline variants | Verifying claims and taking responsibility for errors |
| Finding patterns in large datasets | Investigative work that depends on access and persistence |
| Personalizing newsletters and recommendations | Distinctive voice, analysis, and editorial judgment |
The left column is where AI saves time. The right column is where readers decide whether a publication is worth paying for. A newsroom that cuts the right column to fund the left one is trading its reason to exist for efficiency.
Why publishers are worried anyway
The threat to journalists is less about AI writing better stories and more about AI capturing the audience. When a search engine or chatbot answers the question directly, fewer people click through to the publisher. Radsch's Brookings piece points to the rise of zero-click searches and notes that more than half of 1,159 publishers surveyed had asked AI web crawlers to stop scanning their sites.
Readers are starting to use these tools for news. The Reuters Institute Digital News Report 2025 found that 7% of people use AI chatbots for news each week, rising to 15% of under-25s. The same report says publishers worry AI summaries could further reduce traffic to their sites and apps.
There is also a trust problem. UNRIC describes a 2024 case in which a France 24 journalist was targeted with a deepfake that altered his voice and headline. As synthetic content grows, a recognizable, accountable newsroom becomes more valuable, not less.
Some publishers are turning AI into a revenue line instead. Brookings lists licensing agreements between OpenAI and the Associated Press, Axel Springer, Le Monde, and Prisa. Smaller and local outlets, Radsch warns, are often left out of these deals.
What publishers should do now
Automate the routine work and reinvest the time. Use AI for transcription, tagging, translation, and first-pass summaries, with a human editor approving anything that is published. Set a written policy on disclosure and review before the first AI-assisted story goes live.
Double down on what AI cannot copy. Original reporting, local access, named experts, and a clear editorial voice are what readers pay for. Generic, commodity updates are the content most exposed to AI answers, so they should not carry your business model.
Build direct relationships. Newsletters, registrations, memberships, and subscriptions give you an audience that does not depend on search or chatbot referrals. For more on the traffic side, see how AI Overviews are reshaping digital subscriptions.
How Pelcro helps publishers focus on journalism with AI
Pelcro applies AI to the business side of publishing so newsroom budgets and staff time can go to reporting. Instead of managing billing spreadsheets, support queues, and revenue reports by hand, teams let software handle the operational load.
Pelcro's AI Orchestrator lets teams run the operational layer with plain text commands: create customers, generate invoices, update addresses, and modify subscriptions. Pelcro's AI Customer Service has helped publishers reduce customer service resolution times by approximately 95%, so fewer repetitive tickets reach finance, operations, or editorial staff.
Pelcro's AI Copilot answers plain-language questions about subscriber and revenue data, with no SQL required, and AI Billing is built into the subscription backend. Together they support the strategy that matters most: let AI carry the operational work, and let journalists do the work that cannot be automated.
Frequently Asked Questions
Will AI replace journalists entirely?
No. AI can automate parts of news production, such as routine data stories and transcription, but it cannot do original reporting, build source relationships, or take responsibility for what is published. Researchers at Brookings and the UN's information centre both argue that human journalists remain essential.
What journalism tasks is AI already doing?
AI already transcribes interviews, translates and summarizes documents, generates tags and headlines, personalizes recommendations, and drafts routine stories. The Associated Press began automating corporate earnings reports in 2014, one of the earliest large-scale examples.
What are the risks of using AI to write news?
The main risks are factual errors, bias, and loss of reader trust. CNET corrected 41 of 77 AI-written stories in 2023 after an internal review. Every AI-assisted story should be checked and approved by a human editor before it is published.
Why are publishers worried about AI if it cannot replace reporting?
Because AI can answer readers' questions directly, which reduces visits to publisher sites. Fewer visits mean less ad revenue and fewer subscription conversions. Publishers that rely on commodity content are the most exposed.