After years of enthusiastic coverage, the British broadcast media are starting to be more skeptical about the benefits of AI.
An analysis of more than 215,000 U.K. television and radio broadcasts by PR firm Be Broadcast concluded that AI has entered its “consequence phase”, with a shift in focus away from AI product launches towards accountability, jobs, public services and real-world impact.
Stories are increasingly emerging where AI affects existing public concerns, including child safety, youth unemployment, cyber crime, copyright, energy demand and defense. Meanwhile, planning disputes around data centers are getting more coverage, along with chips, electricity, water and land.
Product announcements are increasingly competing with stories about regulation, legal action, public services and the wider impact of AI, while broadcast coverage is increasingly driven by evidence, independent research, measurable outcomes and credible human voices.
“A year ago the AI story on U.K. television and radio was about products and what they could do for us. Now the conversation covers jobs, energy, copyright and child safety,” commented Stephen Waddington, co-author of AI for Public Relations.
AI is treated very differently depending on the sector. In healthcare and science, it tends to be described as a clinical support tool or route to discovery. There’s more variety in coverage of the use of AI in education, with usefulness and teacher support competing against cheating and assessment integrity. AI coverage of the creative industries, meanwhile, is dominated by ownership, consent and livelihoods.
Cyber security coverage treats AI as both attacker and defender, while government stories position it as a delivery and regulation challenge, and defence frames it as a strategic capability and threat multiplier.
“Broadcasters increasingly want to know who owns the models and who answers when systems cause harm,” said Waddington. “Every organization deploying AI should take note. You will be judged on your position on energy, jobs, safety and rights.”
Productivity is still the most widely-covered AI topic, accounting for 19% of all broadcasts, but now sits alongside investment, copyright, media, transport, employment, healthcare, government and national security.
There’s also been a clear shift in the companies getting the most airtime: while in 2025, coverage centered on products including ChatGPT, Copilot and Gemini, attention has this year moved towards OpenAI, Anthropic, Google and Meta as journalists increasingly focused on ownership, investment, regulation and accountability.
“Broadcasters increasingly want to know who owns the models and who answers when systems cause harm,” said Waddington. “Every positive AI story now attracts a second question about jobs, data, copyright or accountability.”
The shift in angle by broadcasters reflects changes in public attitudes to AI around the world. A recent survey by Ipsos found that while people in Asia and Latin America are on average more likely to think AI has more benefits than drawbacks, in Europe and North America people are on average more likely to feel nervous about it.
In the U.S., previously improving attitudes toward AI have lost momentum, particularly amongst younger people, according to the latest Bentley University-Gallup Business in Society research.
The percentage of Americans who believe AI does more harm than good has increased from 31% in 2025 to 39% in 2026, while the percentage who believe AI does more good than harm has fallen slightly to 9%.
“When public trust is already low, even modest shifts in public opinion like these may matter as businesses continue to expand the use of AI across products, services and workplaces,” the researchers said.
