The speaker

Dr Bahareh Heravi is a Reader in AI and Media at the Surrey Institute for People-Centred AI. Her research is primarily focused on data and computational Journalism, data storytelling, responsible and inclusive AI, information design, and the use of AI in journalism and media. Dr Heravi spoke at a panel on applying AI in small newsrooms at the International Journalism Festival 2024.

The video

Part of our Global Journalism Seminars.

Five takeaways from the talk and the discussion 

1. The arrival of generative AI has popularised AI tools in newsrooms. Whereas AI isn’t new, and AI tools have been used in some newsrooms for up to a decade, the arrival on the scene of OpenAI’s ChatGPT has opened the door for many previously sceptical newsrooms to start thinking about AI, Dr Heravi said. 

“In the past when I was working with news organisations, it was a little hard to sell the idea of AI or computational journalism to them. And it was a very, very slow uptake. Now, everybody's talking about AI,” she said.

2. There are many practical uses of AI tools in the newsroom. There are many ways in which AI tools can simplify journalists’ tasks. Dr Heravi listed several potential applications in many different journalistic areas, including content discovery, investigative analysis, semantic linking, contextualisation, production, storytelling, publishing, distribution, preservation and archiving.

3. There are still some significant limitations. Dr Heravi cautioned against using generative AI tools for producing content relating to current events, due to the ongoing danger of hallucinations, and the fact that many models are still working based on a limited set of training materials and thus are not up to date with breaking news. “If we want to use ChatGPT for its language abilities, it is really good and we should do it. However, we should be very careful and probably not use it for its knowledge abilities, at least when it comes to current events and happenings,” she said.

4. It’s important to address ethical challenges. Journalists have concerns about the use of AI, and with regards to generative AI, they’re particularly worried about lack of human supervision, inaccurate information and bias, Dr Heravi said, referencing a 2024 AP study. 

A follow-up question on the same report asked journalists about strategies for the ethical use of AI. “The highest proportion of journalists said the solution is not using it, which I don't believe is going to be a solution very soon because everybody seems to want or need to use it,” Dr Heravi remarked. 

5. It only takes one interested person to start experimenting. Dr Heravi encouraged journalists, particularly those who work in small newsrooms with fewer resources, to start experimenting with generative AI if they’re interested, and not to wait for more people or additional tools. 

“Start playing with ChatGPT and prompt engineering… Write prompts, read about writing prompts, drop your files in there and see where you get and you'd be surprised, within a month or two you're going to be able to do so much that you would not be able to imagine before starting,” Dr Heravi said. 

The bottom line

The arrival of generative AI has lowered the barrier to entry for AI tools for many within newsrooms, even those with fewer technical skills. There are already many different uses for AI in the newsroom that publications are experimenting with, even small newsrooms with few resources and staff. It’s important to be aware of the limitations of the technology and address ethical challenges that come up, but not using the tools at all is unlikely to be a long-term solution. For those interested, a good way to begin is to start playing with ChatGPT.