TikTok gave teens, children placebo safety feature in experiment, New York alleges

FILE PHOTO: A teenager poses for a picture with a phone displaying the TikTok logo, in Bonn, Germany, February 20, 2026. REUTERS/Jana Rodenbusch/File Photo

Oct 7 (Reuters) - TikTok ran an experiment ⁠on thousands of users who tried to turn on a safety feature on the platform, giving them non-functional versions of the ⁠tool, according to allegations in New York’s lawsuit against the company unsealed on Wednesday.

In the so-called “ghost” or placebo experiments, TikTok selected ‌thousands of users, including teens and children, who opted to use safety features like “Algo Refresh,” a tool designed to reset recommendations for users seeking to escape unwanted or potentially harmful content. According to the lawsuit, those users believed they had activated the feature, but their feeds did not change, unlike users who got the fully working version.

The newly unsealed allegations are ​part of New York Attorney General Letitia James’ lawsuit against the company, one of more ⁠than two dozen brought by states accusing TikTok of designing ⁠its platform to addict children and of misleading users about its safety. Many documents in state cases across the country have been sealed or ⁠heavily ‌redacted as TikTok has sought to keep details about its operations from public view.

The lawsuit claims TikTok assures users that safety is a top priority and that its platform is appropriate for kids and teens, while knowingly preventing some groups of users from accessing certain safety tools, ⁠including users who had specifically sought to avoid unwanted or potentially harmful content.

"We continuously ​make improvements to our platform which means that, ‌like most companies, we regularly test new products and features to validate the experience and understand how they work in the ⁠real world," a TikTok spokesperson ​said in a statement.

The company has broadly denied New York's allegations in court filings and has said that keeping young users safe is a priority.

In September, TikTok settled the first state lawsuit set for trial, agreeing to pay Alabama up to $300 million and change how its platform functions for teens in the state.

James, a Democrat, originally ⁠sued in 2024. The claims about the tests were included in an updated, sealed ​version of New York’s complaint filed at the end of August and later ordered unsealed by a judge.

EMPLOYEE FLAGGED TEST

Internally, a product manager for Digital Well Being at TikTok said doing placebo tests on the Algo Refresh feature was problematic and that users might find it misleading, the lawsuit alleges.

“This feature is ⁠about algorithm transparency and control,” the person wrote in an internal chat, according to the lawsuit. “A placebo group completely conflicts with that.”

Ultimately the company did the tests repeatedly on thousands of users in 2023, comparing users who did not get the feature with those who did to study how long users stayed on the site and the impact on ad revenue, the lawsuit alleges.

The lawsuit also alleges that TikTok misled consumers about how “Algo Refresh” worked, telling ​them it would reset the algorithm that recommended videos to them when it was only a temporary ⁠change.

The unsealed portions of the lawsuit also allege that TikTok conducted a separate set of experiments, which were previously revealed in a report by Bloomberg. In ​them, TikTok randomly selected users and turned offsafety features on their accounts that limited their exposure ‌to harmful content. A teen boy in New York who was part ​of the test committed suicide after he was shown content about self-harm and suicide, Bloomberg reported.

Internal reviews showed the boy had been shown an “onslaught” of content about suicide and self-harm, the lawsuit said.

(Reporting by Diana Novak Jones, Editing by Alexia Garamfalvi and David Gregorio)

Thank you for your report!

Five technology drivers identified as critical to Malaysia's sovereignty, Parliament hears

Technology industry still has room to run

Singapore to strengthen AI safeguards as technology make rapid advances

资料照片：2026年2月20日，在德国波恩，一名青少年手持显示TikTok标志的手机拍照。路透社/Jana Rodenbusch/资料照片

10 月 7 日（路透社）——根据周三公布的纽约针对 TikTok 公司的诉讼中的指控，TikTok 对数千名试图在该平台上启用安全功能的用户进行了一项实验，并向他们提供了该工具的非功能版本。

在所谓的“幽灵”或安慰剂实验中，TikTok 选择了数千名用户，其中包括青少年和儿童，这些用户选择使用“算法刷新”等安全功能。“算法刷新”旨在重置推荐内容，帮助用户摆脱不想要或潜在有害的内容。根据诉讼文件，这些用户以为自己已经激活了该功能，但他们的信息流并没有改变，这与那些使用完整版功能的用户截然不同。

新近解封的指控是纽约州总检察长莱蒂西亚·詹姆斯对TikTok提起的诉讼的一部分。这是二十多个州提起的诉讼之一，这些州指控TikTok设计其平台旨在让儿童上瘾，并误导用户对其安全性存在误解。由于TikTok一直试图阻止公众了解其运营细节，全美各州诉讼中的许多文件已被封存或大量删减。

诉讼称，TikTok 向用户保证安全是首要任务，其平台适合儿童和青少年，但同时却故意阻止某些用户群体访问某些安全工具，其中包括那些特意想要避免接触不想要的或潜在有害内容的用户。

TikTok 发言人在一份声明中表示：“我们不断改进我们的平台，这意味着，像大多数公司一样，我们会定期测试新产品和功能，以验证体验并了解它们在现实世界中的运作方式。”

该公司在法庭文件中对纽约的指控予以了广泛否认，并表示保障年轻用户的安全是首要任务。

9 月，TikTok 与阿拉巴马州达成和解，解决了该州首例即将开庭审理的诉讼，同意向阿拉巴马州支付高达 3 亿美元的赔偿金，并改变其平台在该州青少年中的运作方式。

民主党人詹姆斯最初于 2024 年提起诉讼。有关测试的指控包含在纽约州于 8 月底提交的更新版密封起诉书中，后来法官下令解封。

员工标记测试

诉讼称，TikTok 数字健康部门的一名产品经理在内部表示，对 Algo Refresh 功能进行安慰剂测试是有问题的，用户可能会觉得这种测试具有误导性。

根据诉讼文件，此人在内部聊天记录中写道：“这项功能旨在提高算法的透明度和可控性。而安慰剂组完全违背了这一原则。”

诉讼称，该公司最终在 2023 年对数千名用户反复进行了测试，将未获得该功能的用户与获得该功能的用户进行比较，以研究用户在网站上停留的时间以及对广告收入的影响。

诉讼还指控 TikTok 误导消费者，让他们误以为“算法刷新”会重置向他们推荐视频的算法，而实际上这只是一个临时性的改变。

诉讼文件中未公开的部分还指控TikTok进行了一系列独立的实验，彭博社此前曾报道过这些实验。在这些实验中，TikTok随机选择用户，并关闭了他们账户上的安全功能，这些功能原本可以限制他们接触有害内容。彭博社报道称，一名参与测试的纽约青少年在看到有关自残和自杀的内容后自杀身亡。

诉讼称，内部审查显示，这名男孩曾接触过大量关于自杀和自残的内容。

（Diana Novak Jones报道，Alexia Garamfalvi和David Gregorio编辑）

感谢您的报告！

国会获悉，五项技术驱动因素被认定为对马来西亚主权至关重要。

科技行业仍有发展空间

随着人工智能技术的快速发展，新加坡将加强人工智能安全保障措施。