Q&A: Giving Journalists the Tools to Harness AI

How well-informed is society on function and impact of AI?

Coverage of AI suffers from a mixture of misreporting and underreporting and as a result there are huge literacy gaps. Misleading headlines either exaggerate the benefits of AI or create unnecessary alarm over the threats, including the use of science fiction-style imagery in news reports. We used to say “every reporter should be a climate reporter,” because the subject is simply so vast that it touches every beat and every aspect of social life. Now we say “every reporter should be an AI reporter,” because AI affects everything from health, policing, social security, the environment, and beyond. The victims of AI misuse, particularly in the Global South, may not immediately be able to collectively organize and demand their rights, so journalists everywhere—not only tech reporters—need to be extra alert to their stories.

How does the AI industry respond to this coverage?

Until now AI companies have not been questioned hard enough so they frequently deny interviews, refuse to provide data on the performance of their AI models, and send journalists threatening legal letters. And also there is a disturbing trend of intrusive surveillance software like Pegasus being illicitly deployed against journalists across the world from Mexico to Gaza.

In 2022, the Pulitzer Center launched the AI Accountability Network. What is its purpose?

From our surveys with journalists, we heard an eagerness to take on AI as a subject but also a fearfulness over how intimidating and technical it seems. That’s why we decided to launch the network to equip newsrooms with the training and community building needed to interrogate the technology with confidence. Our year-long fellowships give teams the time, financial freedom, and mentoring by experts from academia to civil society that allow them to investigate the impact of the algorithms that increasingly are part of our lives. Through our cross-border networks, we take a collaborative approach to major issues that no one journalist could do on their own. We follow the money and supply chains of AI that have huge repercussions for the Global South but are controlled in the Global North. For example, our fellows have documented how the AI industry contributes to exploitation when data labeling and content classifying jobs are outsourced for pennies to workers in the Global South; these workers are often unaware that they are contributing to technology used to target dissidents in Russia or China.

How have journalists been able to make an impact using the knowledge gained from AI training?

We have already seen powerful results. In 2023, network journalists working for WIRED and Lighthouse Reports published a series of explosive investigations revealing European governments’ misuse of algorithmic risk assessments against welfare recipients. By forcing public authorities to disclose documents, a multi-disciplinary team successfully audited complex machine learning models to expose how opaque technology was being used to unfairly accuse vulnerable populations of benefit fraud based on discriminatory characteristics such as ethnicity and gender. The outcome was impactful public service journalism that unpacked a flawed tool, while spotlighting human stories. Other teams have uncovered the experimental use of security systems in Greek refugee camps, facial recognition technology in Argentina, and pensioners wrongfully declared dead in India, the latter of which has triggered an Amnesty International investigation.

And how is AI changing the ways in which investigative journalism is produced?

In addition to being a major subject for journalism, there is a major role for AI in journalism, particularly for big, data-driven investigations. The challenge is how to safely incorporate machine learning and other AI subsets into the newsroom. In 2022, fellows from our Rainforest Investigations Network used machine learning and high-resolution and historical satellite images to map for the first time the true scope of gold mining and illegal airstrips in the Amazon rainforest. In cases like this, AI is enabling reporting that would have previously been impossible, or highly time-consuming.

Overall, are you optimistic or pessimistic about the future of AI?

These tools are imperfect and incomplete, and they require constant supervision. By bringing together journalists who cover communities disproportionately affected by AI, we are creating a collaborative ecosystem that empowers media to better report on, and use, this rapidly evolving technology to give audiences a greater understanding on the role it plays in our lives. Recently, the Pulitzer Center launched a wide and ambitious training program for journalists worldwide, prioritizing those in the Global South. The AI Spotlight Series, led by pioneering AI reporter Karen Hao, offers a range of free online workshops and webinars from the introductory to the intensive, as well as a tailor-made course for editors. We should seize the moment and make AI work for us, not the other way round.