New York accuses TikTok of giving teens a fake safety feature that did nothing

New York State filed a lawsuit against TikTok on October 8 alleging the company knowingly gave a placebo version of a safety feature to thousands of users, including minors, while telling them it was working. The case, which cites a Reuters report, centers on a tool called Algo Refresh, designed to let users reset their content recommendations when they're seeing unwanted or potentially harmful material.

According to the allegations, users enrolled in the experiment opted in to Algo Refresh and believed they had turned it on. Their feeds didn't change. A separate group of users outside the experiment, who received a fully functioning version of the same feature, did see their recommendations shift. TikTok, in other words, allegedly ran a controlled test using some users' mental health and safety as the control group, without the fake-feature group's knowledge.

Part of a broader pattern

New York's case is one of more than two dozen state lawsuits alleging TikTok designed its platform to encourage addictive use among children, but this filing adds a specific, documented mechanism to that broader argument: an internal experiment that withheld a safety tool from people who believed they were using it. Bloomberg reported in June on a related case — a separate TikTok experiment that withheld a safety feature from 15 million U.S. users, including a teenager who later died by suicide. That earlier reporting already prompted senators to demand answers from TikTok about experiments that disable safeguards; this lawsuit gives prosecutors a second, more granular example to build a legal case around.

TikTok's response

TikTok told Reuters: "We continuously make improvements to our platform," and said it regularly tests new products and features to understand how they perform in real-world conditions. The company did not respond to TechCrunch's request for comment on the specific allegations in New York's filing.

Why the distinction matters legally

Running A/B tests on product features is standard practice across the tech industry, and on its own isn't illegal or even unusual. What turns this into a legal liability, according to the allegations, is that users in the control group were told — explicitly, through the product's own interface — that a real safety feature had been activated, when it had not. That's the gap between ordinary experimentation and what New York's lawsuit characterizes as deception: not testing whether a safety tool works, but telling vulnerable users it's protecting them while deliberately withholding it.

The case will likely turn on internal TikTok documentation showing what the company knew about the experiment's design and who approved telling users the feature was active. If New York can show executives signed off on presenting a non-functional safety tool as a working one, this moves from a design dispute into the kind of consumer-protection violation that can carry real financial penalties and, more importantly for TikTok's broader legal troubles, sets a documented precedent other state attorneys general can point to.