More than half of newsroom leaders surveyed for a new global study said staff skepticism is the single biggest thing holding back AI integration, as reported in Press Gazette.
The Future Newsrooms Study 2026, from FT Strategies and WAN-IFRA, collected responses from 448 people across 86 countries, most of them editors-in-chief or executive leaders. It found 52% named “cultural resistance or skepticism” as the biggest barrier to wider AI adoption, and 61% blamed a lack of internal technical skills or expertise.
Another 45% pointed to missing strategic direction. Putting the numbers together reveals a pattern: newsrooms aren’t sure what they’re using AI for, staff don’t trust it and nobody’s teaching them how to use it well.
Six in ten newsrooms offer no formal AI training at all. Where training does exist, the report calls it “generic” and “not specific to journalistic needs.”
The study’s central recommendation is structural. Newsrooms that embed an AI authority inside the editorial team, rather than setting strategy from outside, see higher adoption and more confidence among staff. Yet 57% of respondents had no AI expert in the newsroom at all. As the report puts it, having someone “directly within the newsroom to advocate for AI in the editorial context matters to adoption rates.”
Confidence overall is thin. Just 14% of leaders were very or extremely confident their current tech stack was fit for purpose. One in five had no confidence in it whatsoever.
Lisa MacLeod, director of FT Strategies, framed the findings as a warning. Any newsroom “operating on the old playbook of optimising purely for reach and reactive, breaking news, is actively managing its own decline,” she said.
The report also flags a risk in how newsrooms measure success. A majority (43%) expect AI to cut newsroom headcount over the next three years, while newsrooms default to time savings as the main proof AI is working. That, the authors warn, orients newsrooms toward “doing the same work, just faster and with fewer people” instead of enabling journalism that wasn’t possible before.
For publishers, the takeaway is that the AI problem is a management problem. Buying tools solves nothing if editorial staff aren’t brought into the decisions, trained on journalism-specific tasks, and given someone credible inside the room to answer their questions. The tension between how newsrooms talk about AI and how they actually use it shows how fragile trust remains when tools arrive faster than the culture can absorb them.
AI use in newsrooms remains relatively limited. Most organizations use it for transcription and translation (78%), while only 10% have adopted autonomous AI agents, the highest level of agentic AI use, in any newsroom function. Text remains the main focus, with 97% of those surveyed using AI for written content.
Four years after ChatGPT launched, most newsrooms are still using AI as an assistant, not an agent.
“The data in this report—which will be the first of an annual research effort—provides a stark wake-up call,” MacLeod said. “The truth is that our newsrooms are not well prepared for a disrupted future.”
