The BBC is putting generative AI to work in its newsroom—not to replace journalists, but to give them a helping hand. The broadcaster has now rolled out two pilot tools: the handy ‘At a Glance’ summary box and the user-friendly Style Assist editor. Rhodri Talfan Davies, the BBC’s executive sponsor for generative AI, explains that these pilots set out to make news more accessible, with broader rollouts on the horizon if the tests deliver.
After evaluating several in-house AI prototypes, the team zeroed in on two key tasks. First, generating concise, bullet-point summaries that meet the quick-consumption needs of today’s readers, especially younger audiences. Second, reformatting Local Democracy Reporting Service (LDRS) stories so they align seamlessly with the BBC’s style—easing workload backlogs in the process. Davies notes that scannable, short summaries are already striking a chord with many readers.
Over the past 18 months, BBC teams have been trialling various AI tools to tackle these challenges. With the ‘At a Glance’ tool, journalists select from prompt templates, and the in-house language model produces bullet points that are then refined by the reporter. Each summary is clearly marked as ‘AI-assisted and journalist-edited’.
The Style Assist tool works similarly well. It takes incoming LDRS stories, adjusts them to fit the BBC’s tone, flags potential legal issues, and checks for missing attributions—always leaving the final decision to a BBC editor. As Davies points out, nothing gets published without a human final check.
The current six-week pilot will monitor reader engagement with the summary boxes, assess time savings in reformatting LDRS stories, and track error rates following editorial review. Should the results prove positive, we might soon see these tools extended to other areas like national news, sports, and explainer teams.
The BBC’s experience with AI isn’t brand new. Since 2023, it has utilised AI for tasks from transcribing live sports to creating subtitles and aiding in global translations. The focus remains clear: enhance efficiency and accessibility while upholding strict human oversight, in line with emerging ethical guidelines.
This move mirrors broader industry trends. Reuters recently appointed a Newsroom AI Editor, the Associated Press has been automating earnings reports for years, and even The New York Times has integrated AI into its personalised audio app. While AI now supports everything from summarisation to formatting, the editorial decisions still rest with human professionals.
Of course, no system is without its challenges. Early trials revealed some issues with attribution accuracy, prompting more rigorous fact-checking. Concerns about misinformation and legal complexities, especially regarding AI training on private data, remain real. Yet, the BBC’s measured approach shows that technology can support journalists effectively, even when budgets and time are tight.
If you’ve ever wrestled with the pressure of a 24-hour news cycle or struggled to keep up with deadlines, there’s something reassuring about seeing AI tools supporting rather than replacing the human touch. It’s a reminder that while technology can speed up routine tasks, the final call always comes from seasoned professionals.
