Generative AI is rippling through the media industry, but not without raising serious questions
According to Leading Newsrooms in the Age of Generative AI, a new report from the European Broadcasting Union (EBU), artificial intelligence is being embraced by news organisations in a variety of ways—but not without concern, hesitation, and a fair amount of unresolved issues.
The report opens with a flourish: “Do we realise how lucky we are? To be alive in a time as important as the arrival of the printing press, Michelangelo or Leonardo da Vinci.” A bold claim. Whether generative AI earns a place alongside Renaissance masters remains to be seen.
The study does not shy away from outlining major challenges. Among them:
- Business models remain unclear—even for tech firms driving the innovation.
- Many media organisations still haven’t measured a clear return on investment.
- There’s no collective strategy when it comes to negotiating with powerful tech companies.
- Copyright laws around AI-generated content are unresolved.
- Accuracy and hallucination issues persist, making many newsrooms wary of launching AI-facing products.
- AI use is often siloed within organisations; broader newsroom processes have yet to evolve.
- Many outlets, particularly smaller ones, lack formal AI policies altogether.
Despite these roadblocks, the report acknowledges meaningful advances. Automation has improved basic tasks like transcription and translation.
Agentic AI, new model development, and integrations into everyday tools are accelerating how journalists work. In China, the launch of DeepSeek has intensified a global efficiency race.
And adoption is also rising—not just at the institutional level, but among individual reporters adding AI to their personal toolkits.
A separate survey from the Reuters Institute, cited in the report, highlights the most common uses of AI in the media industry:
- Back-end automation (tagging, transcription, copyediting): 60%
- Distribution and recommendation (e.g., personalised homepages): 41%
- Content creation with human oversight (summaries, headlines, visual elements): 30%
- Commercial applications (such as audience segmentation and paywall models): 29%
- Coding and product development: 28%
- Newsgathering and data interrogation: 24%
The report is based on interviews with 20 media leaders and academic experts. EBU members represent nearly 2,000 TV, radio, and online channels, giving the findings broad relevance across the industry.
If there’s a takeaway, it’s that AI in newsrooms is no longer hypothetical—it’s here. But like so many past technological promises, the real challenge lies not in the tools themselves, but in how we use them.
For now, the industry seems to be learning in real time—caught between the lure of innovation and the hard realities of implementation.
