The adoption of AI in newsrooms varies widely: Reuters emphasises speed and automation for market reporting, the BBC prioritises cautious, human-supervised use and clear disclosure, and The Guardian follows a values-driven, reader-funded approach that allows limited, approved AI uses. Reuters uses tools such as Lynx Insight and Fact Genie to publish some alerts in six to eight seconds, while the BBC’s research found AI misrepresented content in 45% of tested cases. Across outlets, journalists face a "verification tax"—the time required to check AI outputs—that shapes whether AI adds net value.
AI Is Reshaping Newsrooms — How Reuters, the BBC and The Guardian Are Taking Different Paths
The BBC has taken a cautious approach to AI, using the technology for specific, low-risk tasks.Lucy North/PA Images via Getty Images
As generative AI tools enter everyday newsroom workflows, leading outlets are adopting very different strategies shaped by their missions, funding models and editorial values. A labour dispute at The Guardian in December 2024 — when nearly 500 journalists struck over the proposed sale of the Guardian-owned Observer — highlighted how those differences can collide with workplace tensions when management reportedly used AI to suggest headlines and write alt text during the walkout.
Three Models of AI Adoption
Reuters: Speed Through Automation
Reuters, a global wire service that supplies news to other publishers, has embedded AI deeply to serve time-sensitive beats such as financial markets. Internally developed systems like Lynx Insight (real-time financial analysis) and Fact Genie (press-release processing) help Reuters publish some breaking market alerts in roughly six to eight seconds — shrinking what was once a race measured in minutes into one measured in seconds.
Automation has expanded coverage and scale, but it raises accuracy challenges. Reuters requires journalists to verify AI-generated claims before publication because automated tools can introduce factual errors. In fast markets, incorrect items are often exposed quickly by trader reactions; in slower-moving investigative or political reporting, mistakes can surface long after publication, complicating risk management.
BBC: Cautious Integration With Strong Oversight
The BBC, funded by the UK licence fee, treats editorial impartiality as central to credibility. It has taken a cautious, tightly governed approach to AI. Primary tools — At a Glance (bullet summaries) and BBC Style Assist (house-style editing) — are used to accelerate routine tasks such as summarising long pieces or reformatting regional stories. The BBC also uses AI translation tools to reach audiences in countries including Poland, Hungary and Romania.
No AI-assisted content is published without human review and sign-off. The BBC published a joint study with the European Broadcasting Union in October 2025 showing AI assistants misrepresented news content — via faulty sourcing, fabricated details or outdated information — in 45% of test cases. The broadcaster publicised the findings and introduced clear disclosure labels: a hexagon icon and the phrase "How we used AI" that appear atop content where AI played a role.
Reuters has embraced AI in its workflow, using it to quickly capture and disseminate financial data.Berkah/Moment via Getty Images
The Guardian: Values-Driven, Reader-Funded Caution
At the time of our research, The Guardian maintained a near-total ban on using generative AI to produce text or images for publication except in narrowly defined cases. In March 2026 it updated that guidance to allow limited generative-AI uses — for example, suggesting alt text, transcribing audio or analysing parliamentary papers — provided the tools meet The Guardian's editorial standards, remain under human oversight and have explicit approval from a senior editor.
“Our approach to generative AI is designed to support our journalists’ expertise, never replace it,” The Guardian told The Conversation. “Use of these tools requires absolute rigour and responsibility.”
The Guardian’s trust-based, reader-funded ownership model (established in 1936) helps insulate the paper from short-term commercial pressures and explains its cautious stance. During the December 2024 strike, many staff viewed management's deployment of AI as undermining core principles and labour rights, underscoring tensions that can arise when operational pressures meet organisational values.
Common Challenges and The Verification Tax
Across organisations, one shared cost is what the authors call a "verification tax": the additional time reporters and editors must spend checking and correcting AI outputs. If staff spend more time fixing chatbot errors than reporting original stories, the technology's net value can shrink.
We argue that AI will not transform the industry uniformly. Commercial outlets are likely to increase automation to scale coverage; public broadcasters will emphasise human oversight and transparency; and independent, reader-funded outlets will prioritise journalism generated and verified by people. The most effective newsrooms will balance speed with accuracy, treat staff fairly and protect the trust that underpins quality journalism.
Authors: Erik P. Bucy and Milad Jalalian Ebrahimi, Texas Tech University. This article was republished from The Conversation, a nonprofit independent news organisation.
Help us improve.
