Artificial intelligence is reshaping how people do their jobs, and for some newsrooms that means using AI as a tool to strengthen their reporting.
Alex Mahadevan is the director of MediaWise and on faculty at the Poynter Institute — a nonprofit media institute and newsroom that trains journalists and the public on media literacy and journalism ethics. He also leads Poynter’s AI Innovation Lab.
Part of Mahadevan’s job is to research ways AI is being used to inform and misinform people. He also trains journalists to understand AI tools and use them ethically to build audience trust.
A recent Lake Effect special explored how AI is impacting different parts of our workforce. Mahadevan spoke with Lake Effect’s Xcaret Nuñez about how journalists are grappling with using AI.
This conversation has been edited for length and clarity.
Xcaret Nuñez: A big part of your job is training newsrooms and journalists on how to use AI ethically and responsibly. What are some reasons journalists you've talked to are hesitant about trying AI tools in their work?
Alex Mahadevan: Journalists are generally skeptical of AI tools because I think as reporters, we are generally skeptical of big tech. These are tools that have come out of Silicon Valley and are closely associated with some of the same folks who brought us the social platforms that completely changed how we reach audiences and have pulled the rug out from under journalism before. So there is a hesitancy there just because this is a new technology put out there by people that we sometimes can't really trust.
The second thing is how this technology was developed. These large language models were trained on our reporting. I mean, some of my reporting was scraped from the internet and put into these tools. Now, Poynter doesn't charge for its content, but for news organizations that do, I think they would see that as theft. So I think there is that concern about using a tool that was trained on our own content that we worked really hard on.
And then lastly, there is definitely a concern about the environmental impact. These tools use a lot of energy, which puts pressure on our cloud infrastructure, which has led to the development of data centers, which obviously have become a huge story for local residents who are starting to see these pop up in their backyards. So I totally see why reporters are questioning these tools, even though it is my job to make sure they use them responsibly and see a lot of opportunity in using these tools to do better work and reach more people.
What does it mean to use AI ethically in the context of journalism?
I would say using AI ethically just means using it in a way that aligns with your organization's existing editorial policies and your own moral and ethical guidelines. I do think every news organization needs a separate AI policy that generally mirrors what is in their own editorial policy and is built around principles that align with our own journalistic principles. So, the other reason that journalists are very skeptical is because big tech does not share our values. Our values are around accuracy, service and accountability. They are about overcoming bias and making sure all voices are heard.
Now, when you think about big tech's values, a lot of these tools are opaque, so they are not transparent or accountable. There is not a lot of thought about the bias in some of these machines that are being generated. These tools produce errors because they are probability machines. So when you use them to do data analysis, double-check it and make sure it's accurate.
If you are using it, do something that would have made a story not possible. So let's say you’re sifting through lots and lots of campaign finance documents, and you find some really cool stories to write out of that and you write them up. You would disclose that as “This tool was used to do this analysis. A human reviewed this and this is why we did it.” So you are transparent about your AI use, and overall using it in a way that's efficient. I teach journalists how to use these tools the least amount possible. So if you use these tools really effectively, you are also reducing the load you put on the cloud infrastructure.
Can you share an example of a newsroom that's testing AI tools successfully?
My hometown newspaper, the Tampa Bay Times, uses AI for weather alerts and for real estate briefs. These are two pretty high-demand article types that people need in our area, and they're something that took a while to put together and took journalists off of bigger stories. But we recognize they're important. So now they have automated those two things, so journalists, instead of spending so much time writing real estate briefs, they can focus on bigger investigations on real estate. The Tampa Bay Times did a great investigation on hedge funds buying up homes in my neighborhood and other neighborhoods. Investigations like that would not be possible if journalists had to do some of the rote and boring things they had to do in the past. It's more time to do investigations.
It seems like you believe newsrooms have to experiment with using AI tools to survive, or is it possible for a newsroom to stick to its ways?
There are very specific brands that can absolutely flourish and do great without using AI. But I think traditional news outlets who want to do really good investigative work and great data journalism are doing a huge disservice to themselves [by not using AI tools] — and quite honestly, their audiences — if they are actively avoiding using something that can find secrets in campaign finance records and can dig through a thousand emails from your county commissioners and figure out when they've been consorting with developers on a new data center. There can be great reporting on data centers that can happen a lot quicker and more in-depth with using these AI tools. So I'm very passionate about traditional news organizations experimenting, but there are definitely brands who are better off not doing it and have no use for it. It is just a tool after all.
You can learn more about AI literacy and ethics at the Poynter Institute and MediaWise’s website.
_
