TikTok Ran the Experiment. Chase Nasca Was in the Control Group.

In December 2021, TikTok issued a press release announcing it was testing ways to "interrupt repetitive patterns" of harmful content. What it left out: when TikTok tweaked its algorithm in 2021 to stop users from being overwhelmed with harmful content, the company didn't roll out the safer version to everyone. Instead, to see if the change might reduce the app's stickiness, it conducted an experiment.

That is not a safety story. That is a product management decision with a body count.

TikTok created a control group of 10 percent of US users — at the time, roughly 15 million people — who kept the old version of the app. This group included 16-year-old Chase Nasca. TikTok's algorithm pushed him thousands of videos about suicide, sadness, hopelessness and loneliness, right up until he killed himself.

Translation: TikTok had a safer algorithm. It chose not to give it to 15 million people. It wanted to know how much engagement the dangerous one was worth. Someone died while the meter was running.

What the Document Actually Changes

The prior legal posture: we did not know the specific harm, we cannot curate every video, Section 230 protects our editorial decisions. The Nasca family has been in litigation since 2023.

The lawsuit alleges that TikTok purposefully sent Chase more than 1,000 videos promoting suicide, hopelessness and self-harm in order to maximize his level of engagement on the platform resulting in higher ad revenues.

TikTok's response was Section 230 protection from liability for third-party content.

Bloomberg's document destroys that foundation. This is not about which videos third parties uploaded. The document explains that Chase was fed this content because his account had been randomly selected for an "internal backtest," making his account part of the control group. For him, TikTok's safeguard wasn't switched on.

A backtest is not curation of third-party content. It is a deliberate product choice about which version of TikTok's own system a user receives. Section 230 does not protect a company from its own product decisions. This distinction is the entire ballgame.

The day before Chase died, he scrolled TikTok for more than five hours and searched for posts about suicide and trains. On the day of his death, he favorited videos alluding to suicide. Hours before he died, he watched and rewatched a video about a teen boy who left a goodbye message on social media before jumping from the Manhattan Bridge. The last video Chase saw was of someone crying.

The algorithm that could have interrupted that sequence existed. By early 2022, TikTok had rolled out its newly safeguarded algorithm to 90 percent of US users. It held back 10 percent in the control group.

What Every Other Platform's Board Reads This Morning

The liability theory emerging from this document is not algorithmic negligence—courts have struggled to define it. It is conscious withholding of a safer product from a captive user base to measure engagement uplift. That is a product liability framing. It is a consumer protection framing. It is the framing state AGs in 14 states have been hunting for.

If your company has internal documentation showing you tested a safer version of a feature and measured what it cost you in time-on-app, you have a document a plaintiff's attorney will now subpoena.

The experiment itself is not obviously illegal. What transforms it into liability is the combination: foreseeable harm, a known safer alternative, and a documented decision to withhold it from a randomly selected population.

That is not an accident. That is a clinical trial on a minor without consent.

TikTok has not yet contested the document's authenticity on the record available at publication. Whether it does determines whether this is a litigation event or a legislative event. Either way, the document exists.

What to Watch

• Whether TikTok disputes the document's authenticity or argues the experiment was mischaracterized — that response lands within 72 hours or not at all.
• Congressional hearings: the Nasca family attended a House committee hearing in March 2023 and will likely return within two weeks.
• Discovery requests hitting Meta, YouTube, Snapchat, and Discord: any internal document showing a safer feature was tested against an engagement baseline now carries legal exposure.
• Advertiser response, specifically brands with public brand-safety commitments.
• Whether a TikTok employee files a whistleblower complaint with additional documents.