Methodologies for tracking and reporting AI
Hilke Schellmann is an Emmy-award-winning journalism professor at New York University and a freelance reporter covering artificial intelligence. Her work has been published in The Wall Street Journal, The Guardian, The New York Times, and MIT Technology Review. Her recently published book, The Algorithm, explores artificial intelligence and the future of work.
Part of our Global Journalism Seminar series.
Five takeaways from the talk and the discussion
1. Make sure your reporting is balanced. Schellmann started by warning against getting caught up in both the hype and doom around AI. For journalists, she suggested, it can perhaps be tempting to lean into the excitement and see it as a positive thing that will drive readers to books and articles. However, what will lead to valuable reporting is to look into the impact of existing AI tools in various industries. “As we all know, there's a lot of hype in the space. And journalists sometimes fall into that trap, too… But I'm trying to really get away from this, and I get this asked all the time, do I think that technology is going to solve all our problems or am I a doomsday scenario kind of person? And neither, what I want to do is really look into what is already happening,” Schellmann said.
2. Cultivate insider sources. Schellmann explained how she used insider sources, including employment lawyers and industrial-organisational psychologists, to gather information on AI-powered hiring tools for an investigation. “My background is being an investigative journalist. So I've been trained in investigative methods. So I try to use some of them, such as trying to find insiders willing to speak,” Schellmann said. Other helpful sources have been ex-employees of companies in the field who were open to sharing their experiences.
3. Go to conferences. “I also went to a tonne of conferences, and I love conferences because so many people just want to chat even though I have a badge, they can clearly tell I'm a journalist. Sometimes I get to meet with CEOs because they were at the booth so maybe they don't answer my PR inquiry when I call them or I LinkedIn message them but they're actually there,” Schellmann said. As well as talking to industry insiders, conferences are also useful for accessing product demos, she added.
4. Be sceptical. Schellmann emphasised the importance of scepticism when evaluating AI tools. “I urge journalists to be very sceptical and ask for a clear explanation. Ask if developers can clearly, in layman's terms, explain,” she said, adding that the normal journalistic scepticism applied to other fields seems sometimes to be missing in AI-focused reporting. "For some reason, we as journalists aren't too sceptical, I think because it feels so complex, and there aren't a lot of other folks who can put things into perspective. So when a company says, it's 90% accurate, we just take that."
5. Work with academics. Journalists and academics can work together symbiotically to help each other explore and better understand the field of AI. “I think it could be helpful to be like a journalism co-pilot with them to understand, what are they finding out, can I report on that and help them. I actually feel like there's something really fruitful that we as journalists can bring, we have our ears to the ground to understand, okay, this is happening. Researchers don't actually know what is happening inside companies, inside healthcare, the quick changes. And we can help them have good questions to ask to get the right data,” Schellmann said.
The bottom line
The field of AI is fast developing and can sometimes appear obscure due to a lack of transparency, overuse of technical jargon and few established experts. However, journalists shouldn’t let themselves be intimidated and instead maintain their scepticism, push for clear explanations and delve into the world of existing technologies instead of sticking to hype/doom narratives. They can apply existing journalistic skills, such as investigative methods, as well as develop innovative ways to report on AI, for example through collaboration with researchers.