Inside the documents that could break Big Tech's teen-safety defence
Internal records unsealed in the first US bellwether trials suggest Meta and Google knew the harm their platforms did to teenagers years before they shipped the features. Section 230 may not save them.
What it’s about
The first major social media liability trials are now under way in the United States. Verdicts in California and New Mexico have ordered Meta and Google to pay hundreds of millions of dollars in damages over the past few weeks. A Handelsblatt investigation shows what’s in the case files: internal records and chat logs suggest the companies knew for years about the dangers their platforms posed to teenagers – and kept their business model running regardless. For the industry, this is a tipping point.
The Cusato case
Gabriella Cusato was 15 when she hanged herself in her closet in November 2019, after her mother had taken her phone away following an argument. Eating disorders, several secret Instagram accounts, algorithms that – her family is convinced – pulled her ever deeper into a “rabbit hole” of thinness, beauty and self-hatred content. The Cusatos are now suing Meta together with around 1,800 other families. The Californian case is being treated as a bellwether trial – a test case that will set the direction for the next 2,000 lawsuits.
What the documents show
Meta plans to appeal and accuses the prosecution of using “selected quotes” and a “misleading presentation”. Google takes a similar line. But the internal documents that have become public through the trials paint a different picture: studies on suicide risk, on eating disorders, on cognitive dependency were conducted internally – and the findings were on the table before products like Reels or infinite scroll were rolled out. “The past is catching up with us,” says one Meta employee in Silicon Valley.
Why the case is bigger than it looks
The trials are a litmus test for the liability question around digital platforms. Until now, Section 230 has shielded US platforms: what users post is not the platform’s responsibility. Product design, however, is a different matter. If a court decides that algorithms themselves are a product that causes harm, the same liability logic that applies to medical devices or cars kicks in. Europe is watching closely. The Digital Services Act points in the same direction, but more cautiously so far.
I reported the full story together with Felix Holtermann, Josefine Fokuhl, Thomas Jahn, Kevin Knitterscheidt, Gunter Nowy and Olga Scheer for Handelsblatt. This is a first impression.