Analysis
New York's attorney general alleges TikTok ran an experiment that gave some teens and children a non-working version of a safety feature called Algo Refresh, while leading them to believe it was active, according to a new state lawsuit reported by TechCrunch.
What the lawsuit claims
Algo Refresh is meant to let users reset their recommendations and stop seeing unwanted or potentially harmful content. The lawsuit alleges that in an experiment involving thousands of users, participants who opted into Algo Refresh believed the feature was active, but their feeds did not actually change -- while users outside the experiment group received a fully functioning version of the same tool. New York's suit also alleges TikTok ran additional, undisclosed experiments beyond this one.
“## What the lawsuit claims Algo Refresh is meant to let users reset their recommendations and stop seeing unwanted or potentially harmful content.”
Part of a broader, escalating legal pattern
New York's case is one of more than two dozen state lawsuits accusing TikTok of designing its platform to encourage addictive use among children and of misleading users about its safety tools -- the same category of claim behind the $100M Alabama settlement Pulse covered last month. It follows a June Bloomberg report that a separate TikTok experiment withheld a safety feature entirely from 15 million US users -- including a teenager who later died by suicide -- a case serious enough that US senators publicly demanded answers from the company. Taken together, the two reported experiments suggest TikTok has repeatedly tested safety-feature access on real users without their knowledge, not as an isolated incident.
TikTok's response
TikTok told Reuters it regularly tests new products and features to validate the experience and understand how they work in the real world; the company did not immediately respond to TechCrunch's request for comment on this specific lawsuit.
The counterweight
TikTok's defense -- that split-testing features is normal product practice -- is true as a general matter; however, the specific allegation here is not that a feature rolled out slowly, but that users were told a safety tool was protecting them when, per the complaint, it was not. Whether courts treat that distinction as a meaningful legal line or as ordinary A/B testing that happened to involve a sensitive feature will shape how every consumer platform runs similar experiments going forward. The suit's claims have not been proven in court.
What to watch
Whether other states amend their existing TikTok suits to add similar placebo-feature claims, and whether this becomes the test case that determines if split-testing a safety feature, specifically, requires disclosure that ordinary feature testing does not.
