Is artificial intelligence killing journalism – or saving it? As large language models continue to infiltrate newsrooms, fears of job losses, misinformation, and editorial decline have taken centre stage.
Behind the alarm bells, a quieter reality is unfolding. While the public braces for disruption, journalists are harnessing AI to break stories, expose the truth, and push investigative reporting into uncharted territory.
Two recent studies – one measuring AI’s actual labour impact, the other capturing public anxiety – reveal a striking disconnect between perception and practice.
Real vs perceived impact
A recent working paper titled “Large Language Models, Small Labour Market Effects” by economists Anders Humlum and Emilie Vestergaard argues that the rapid emergence of generative AI tools has so far had only a modest impact on employment levels in journalism and related professions.
Using Danish labour market data and models forecasting occupational exposure to LLMs, the study found no significant decline in job numbers or hiring rates, even in sectors most exposed to these technologies. Rather than triggering mass layoffs, AI appears to have been integrated into workflows without displacing journalists en masse.
This empirical finding stands in contrast to widespread public scepticism. According to a Pew Research Centre survey conducted in the summer of 2024, half of US adults believe AI will negatively affect the quality of news over the next two decades. Just one in ten expects a positive impact.
Concerns are particularly strong around employment: nearly 60 per cent of respondents think AI will lead to fewer jobs for journalists. Trust in AI-generated content is also low, with two-thirds expressing deep concern about the spread of inaccurate or misleading information.
Perhaps surprisingly, this mistrust cuts across political affiliations.
In the United State, both Democrats and Republicans share similar concerns about misinformation and potential job losses, suggesting that the anxiety is not merely partisan, but structural.
Education appears to play a greater role. College-educated respondents are more likely to view AI as a threat to news quality and job security than those with less formal schooling.
Using AI to enhance reporting
Yet, even as the public wrestles with the risks of AI, many journalists are quietly proving its value.
A Nieman Labs analysis looked at several Pulitzer Prize finalists – including one award winner – where applicants were required to disclose their use of AI. in these cases, they did not rely on generative models, but on earlier machine learning techniques and custom-built algorithms, used as essential tools in investigative work.
The Wall Street Journal used machine learning to analyse over 41,000 of Elon Musk’s posts on X (formerly Twitter), visualising his political transformation.
The Washington Post applied geospatial AI to satellite imagery to challenge the Israeli military’s account of a drone strike that killed two journalists in Gaza.
Meanwhile, the Associated Press and partners built a new database of deaths caused by “less-lethal” police force, using AI-driven document analysis and transcription tools.
Another finalist project, 40 Acres and a Lie, used image recognition to search 1.8 million historical documents and uncover the stories of formerly enslaved people who were stripped of land after the US Civil War.
Adaptation, not displacement
Crucially, in all these cases, AI served as a tool, not a substitute. Reporters used it to expand the scope of their investigations, accelerate document review, and enhance visual analysis. Human judgment, fact-checking and narrative storytelling remained central to the work.
As Pulitzer administrator Marjorie Miller put it, AI – when used responsibly – adds “agility, depth and rigour” to journalism, but does not replace it.
These insights suggest that the future of journalism in the age of AI may be defined less by displacement than by adaptation.
While concerns about misinformation and automation are valid and require continued scrutiny, the growing number of journalists using AI constructively paints a more nuanced picture – one in which technology augments human expertise, rather than eliminating it.
Perhaps in the debate over AI’s role in the media, it is time to listen not just to public fears or tech-industry hype, but to the reporters already using these tools to tell stories that matter.
[Edited By Brian Maguire | Euractiv’s Advocacy Lab ]
