How to Prepare for the AI Whistleblowers
In the absence of better laws, journalists can still make whistleblowing less painful
In the summer of 2021, my job was reporting on Facebook. Seemingly every week, there was a new story about how the company favored growth over safety, allowing conspiracy theories, hate speech, and worse to flourish. Everybody could see it happening from the outside. Surely, I thought, someone on the inside would deliver some hard evidence sooner or later.
Then, in October, that whistleblower arrived. Frances Haugen, who was employed on a safety team at Facebook, leaked thousands of internal documents to the Wall Street Journal, revealing the extent of what Facebook knew about the problems on its platform, and how, in many cases, cold business reasons prevented it from addressing them.
We haven’t yet had a big whistleblower from a “frontier” AI lab, but right now I feel the way I felt that summer before Haugen went public. AI companies, fuelled by big tech money, are racing one another to build bigger and more capable AI models. Meanwhile, there are many workers inside these companies who care deeply about ethics and safety — and who worry intensely about the negative externalities of what they see as dangerous industry “race dynamics.” It’s clearer than ever that the business side of AI and the safety side of AI are in a shaky alliance. There are plenty of disincentives to whistleblowing: intimidating NDAs, the risk of ostracism, the harsh glare of public attention. But at some point, someone will decide the risks of inaction are so high that speaking out is worth the pain.
How can journalists help minimize that pain? A couple of weeks ago, I discussed this question on a panel in Italy with Haugen and Tyler Shultz (who blew the whistle on the fraudulent blood testing startup Theranos). The panel, titled How to Prepare for the AI Whistleblowers, was organized by the Signals Network, a whistleblower protection group that we’ve all worked with in the past. The Signals Network supported my work on a story about Facebook's African content moderators, giving independent legal advice and post-publication psychological support to my source Daniel Motaung. They’ve offered similar help to Haugen and Schultz, not only supporting them in getting their messages out in public, but also providing crucial pastoral care during the lonely, isolating task of becoming a public whistleblower. In other cases, they’ve provided sources with temporary safe houses and secure communications technology.
On the panel, we discussed the disincentives stacked against whistleblowers, and the measures that journalists can take, in the absence of better whistleblower protection laws, to look after their sources better. Tyler and Frances shared lessons they’ve learned. Delphine and Jennifer (from the Signals Network) shared crucial advice for both potential whistleblowers and the journalists working with them. And I shared some details of my work with the Signals Network in the past, which I think is a blueprint for how journalists can do a better job looking after whistleblowers before, during, and after publication.
The full panel is available to watch on YouTube, here:
As always, my Signal username is billyperrigo.01 if you want to get in touch.