On this week’s episode of Media Voices we hear from Tshepo Tshabalala, project manager & team lead at the LSE’s JournalismAI project. Rather than focusing on the negative coverage and speculation we so often hear about AI replacing journalism jobs, Tshepo tells us about uses cases and the benefits of AI in smaller newsrooms and how collaboration is helping journalists get on board the AI rocket ship.
Tshepo joined the JournalismAI project just three months after ChatGPT went public. Some might call that perfect timing; for others it would be a nightmare, with the playing field changing week by week. He’s taken it all in his stride, however, telling us, “It was a rollercoaster… trying to learn the job, trying to understand the field and understand who the players are, also serving our audience and our users that need help. But growth doesn’t happen in a comfortable space.”
Knowledge-sharing is at the heart of everything the JournalismAI project does, from the JournalismAI Starter Pack, designed to help news organisations understand the opportunities offered by AI, to the 2023 JournalismAI report surveying 105 newsrooms in 46 countries. Tshepo says, “It’s really just sharing knowledge, and then with the hope that they can decide thereafter, what they do with that knowledge.”
In the news roundup the team discusses Sydney Sweeney, at length. They also discuss the disparity between how much journalists and creators get paid versus the shareholders of media companies, the news that GB News made an enormous loss last year, and ask whether it’s good or bad for TalkTV to be going digital-only on YouTube.
Some highlights from the interview, lightly edited for clarity:
Key takeaways from JournalismAI’s latest report
In this new report, we interviewed 105 newsrooms from 46 countries. We did in-depth interviews with all of them. So for the first time, we really made sure that we had representation from South America, Africa, Asia. We had about 20-24% of newsrooms from each region of the world.
The biggest takeaways from that report in a nutshell is that we learned that these rapid experimentations with AI, especially with generative AI… we were learning that a lot of newsrooms were starting to experiment with generative AI. We learned that there was a key concern about the ethical integration of AI, and also a question about upholding journalistic values within the newsrooms, which is the challenge a lot of organisations are facing.
The beauty of making sure there was good representation, we learned that AI adoption is unevenly distributed across the world. Adoption is a lot higher in the Western markets, and a lot lower when you go to South America, Asia and Africa.
Lastly, something we saw in a previous report that came through again this time around is the need for transparency, both from the tech companies but also from news organisations in how we use AI. Second to that is collaboration; the need to collaborate, learn from each other, and growing the literacy of AI amongst journalists, as well as newsrooms.
Using AI in newsrooms
From the research we learned that there were some people who, for example, were using AI platforms – mostly generative AI platforms – for themselves. The newsrooms hadn’t had a policy yet on the usage of AI. So you’d have [someone] there who sits at his desk and uses ChatGPT to refine a headline, or whatever. So in that way, they use that for themselves to help them with their work, even though there wasn’t a policy in place already.
Some news organisations had guidelines in place already, and they were able to guide the journalist in how to use AI in the newsroom using the various technologies that they had adopted already. So it really varies based on the respective newsrooms.
I think the difference is, what we learned is that smaller newsrooms have bigger room to experiment, whereas the larger, traditional newsrooms don’t have that space. They’re more traditional, more bureaucratic. So it’s a lot more difficult to even think about experimenting or using AI, whereas other newsrooms that are bigger or smaller, where they have thought of or started using AI, it’s a lot easier, the more nimble and flexible in how they adopt or think about using AI.
Optimism about the future of AI
I’m always optimistic. Once you understand that, with anything, there will be those who use it for good and those who use it for bad, I’m always looking at the good side to say listen, yes this is for good, but in the back of your mind, know that you’re challenging or working against others who are using it for bad. There’ll always be this back and forth between the two.
I think a lot of people will try to figure out what’s the best-case scenario. Obviously the newsrooms who already have implemented or used LLM or language models within the newsrooms already, but generative AI has brought in a new element that wasn’t there before. So in the next couple of months to a year, we’ll get to see use cases of how newsrooms are actually using it.
The news round-up
Whatever your view on the wealth inequalities in today’s society, a Defector story about wealth inequality in Hollywood by Kelsey McKinney is a very pertinent look at some of the problems bubbling under the surface in not just film, but publishing too. Some pertinent quotes:
- “Ernest Hemingway was paid $1 a word in 1936. That’s more than $21 per word in today’s dollars,” McKinney writes. “No one (and I really mean no one) in media makes $21/word. That compensation just doesn’t exist. You could be the most popular novelist in the world and not make $21/word to report. You could argue that no writer today is as good or popular as Hemingway was at his peak, but no writer today is even making half or a quarter of what he made, and writers only ever get so famous. If someone were paid $5/word in 2022…that would be a quarter of what Hemingway was paid. That writer would be able to pay their rent and health insurance premiums and tuck some money away in savings off a standard-issue story per month, but again, that lucky writer does not exist.”
- “Why should any CEO make more than the actresses whose labor and beauty they sell? Why should a second-year management consultant at every major consulting firm make more than every single writer I have ever known? It’s not even a question of principle. People buy things: services and products and experiences and feelings. How is it that the creation and provision of those things is valued so little, when it is so essential?”
- “The reality is that the people with the most money have devised, at every turn, new and more bulletproof ways for them to make and keep more money, and for the people who make things to make less. This is the eternal story of labor and management; it just has hot people in it, in this case.”
In mid February, The National Association of Press Agencies (NAPA) warned of agency rebellion over some freelance rates from News UK (Sun and Times) that hadn’t changed in 40 years.
A report from the Reuters Institute in January spoke to 10 young journalists about their struggles to break into the news industry, citing low salaries, long hours and uninspiring roles that never actually lead to full-time jobs.
Earlier in the week GB News announced losses of £78 million on the same day that TalkTV gave up on linear broadcasting.
