Two juries delivered guilty verdicts against Meta this week, and the fallout reaches far beyond social media. As reported by Hacker News, juries in New Mexico and Los Angeles found that Meta failed to protect kids on its platforms, relying heavily on the company’s own internal research as evidence.
The core problem: Meta hired researchers to study how its products affected users, then tried to bury the findings when they turned ugly. Internal surveys showed alarming rates of teenage users receiving unwanted sexual advances on Instagram. Other research implied that people who cut back on Facebook became less depressed and anxious. Meta eventually halted that study.
Brian Boland, a former Facebook executive who testified in both trials, put it plainly: the internal findings contradicted how Meta presented itself publicly. “Both juries, with very different cases, came back with clear verdicts,” he said.
Why AI Companies Should Pay Attention
This is where the story gets interesting for the broader tech industry. After Frances Haugen’s 2021 whistleblower leaks exposed Meta’s internal knowledge of product harms, newer AI companies like OpenAI and Anthropic invested heavily in safety research teams. They hired researchers, published findings, and positioned themselves as the responsible alternative.
Now those companies face an uncomfortable question: does funding internal safety research create legal liability down the road?
Meta’s response after Haugen was to clamp down on research teams. Many groups studying platform harms were cut. The company started removing tools that third-party researchers used to study its platforms. Other tech companies followed the same playbook.
The Research Suppression Trap
Lisa Strohman, a psychologist and attorney who consulted on the New Mexico suit, identified the miscalculation: “I think what they failed to recognize is that researchers are parents and family members. These people weren’t going to be bought.”
Kate Blocker, director of research at the nonprofit Children and Screens, raised the alarm about a pattern repeating itself with AI: “Much like with social media before it, there is limited public visibility into what AI companies are studying about their products.”
She noted that AI companies appear mostly focused on studying models themselves (behavior, interpretability, alignment) rather than the real-world effects on users. That’s a significant blind spot.
What This Means Going Forward
The Meta verdicts create a paradox for AI companies:
- Fund research and risk liability. If internal studies reveal harms and the company doesn’t act, those documents become courtroom ammunition.
- Suppress research and risk worse outcomes. Without studying harms, companies can’t fix them. And when problems surface anyway (they always do), the suppression itself becomes evidence of negligence.
- The third option: transparency. Companies that publish findings and act on them have the strongest legal and ethical position. But that requires genuine commitment, not PR strategy.
Sacha Haworth, executive director of the Tech Oversight Project, noted that most of the internal research used in this week’s trials wasn’t new. What mattered were “the very emails, the very words, the very screenshots” that showed executives knew about harms and chose inaction.
Meta and YouTube (also a defendant in the L.A. trial) said they would appeal. Meta’s defense argued the research was old, taken out of context, and misleading.
But the signal to AI companies is clear. The industry is pushing aggressively into consumer AI products while deprioritizing safety research. If these verdicts hold, the legal playbook for holding AI companies accountable just got a major upgrade.
The full story is available at the original source for those who want the complete details.