发生了什么
Startup Fortune reports that an amended class-action complaint filed August 31 in the U.S. District Court for the Northern District of California accuses Meta of routing footage and audio captured by Ray-Ban Meta glasses to Kenya-based Sama contractors for labeling and possible AI training. The amended complaint reportedly expands the proposed classes beyond glasses purchasers to include bystanders who were recorded without agreeing to Meta’s policies.
Startup Fortune says the amended complaint is sharper than the original March 4 lawsuit, which focused on purchasers Gina Bartone and Mateo Canu. The article reports that the proposed case now also covers bystanders whose faces, voices, homes, bedrooms, children, or other private settings may have been captured by someone else’s glasses. According to the article, the complaint alleges that imagery and audio could be sent to Meta’s servers for analysis and possible model training.
Startup Fortune also reports that a February investigation by Svenska Dagbladet and Göteborgs-Posten, conducted with Kenyan journalist Naipanoi Lepapa, found that Sama workers reviewed Ray-Ban Meta footage containing nudity, sex, bathroom use, banking details, and other private material. The article says the investigation drew on interviews with more than 30 Sama employees and that Meta ended its contract with Sama afterward. These underlying allegations and the complaint’s claims were not independently confirmed here.
为什么这很重要
The reported allegations concern a direct link between consumer wearables, AI training data, outsourced human review, and people who never chose to use the product. If substantiated, the case could raise significant questions about consent, notice, data handling, and responsibility for recordings captured in homes and public spaces. The allegations remain unproven, and Startup Fortune reports that Meta had not publicly responded to the amended complaint.
The case focuses on a practical privacy problem created by AI-enabled cameras: the wearer may have interacted with Meta’s product terms, but a person recorded nearby may not have received notice or had a meaningful opportunity to object. The reported use of contractors adds another layer, because sensitive recordings may be viewed and labeled by workers outside the country where they were captured.
The source also places the lawsuit in the context of Meta’s broader privacy and child-safety disputes, but those matters involve different allegations and should not be treated as evidence that the smart-glasses claims are true. The article’s account does not establish how much footage was collected, whether every category of footage was used for training, or whether any model was trained on specific recordings.
互动机制：它实际上是如何运作的
以交互方式探索这一发展背后的基础技术。
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
接下来看什么
Watch for Meta’s response, court rulings on the proposed classes, and evidence about what footage was collected, how it was used, and whether it entered model-training workflows. The source does not document product access conditions or pricing changes.
The next important developments are Meta’s public response, any motion to dismiss or challenge to the proposed bystander classes, and judicial findings about consent, notice, and the company’s data practices. A lawsuit filing establishes allegations, not liability.
Further reporting or court records could clarify the Sama contract’s scope, the review and retention process, the identity of any AI systems involved, and whether Meta’s privacy controls changed after the reported investigation. Startup Fortune reports no confirmed access or pricing changes for the glasses in connection with this case.