Respecting Privacy: Amazon's $25M Settlement with FTC Over Kids' Privacy and Deletion Requests
Amazon has agreed to pay a $25 million penalty to the Federal Trade Commission (FTC) and make changes to its privacy practices to settle charges of violating the Children's Online Privacy Protection Act (COPPA). The FTC accused Amazon of retaining children's voice recordings indefinitely and failing to delete the data even when requested by parents. The company also kept transcripts of children's speech in its databases and did not delete geolocation data, despite being aware of these practices since 2018.
The FTC relaxed COPPA rules to accommodate reasonable usage of voice assistants like Amazon's Alexa by children. However, Amazon's retention of children's speech data for extended periods went beyond what was considered reasonably necessary. The company claimed that the data was crucial for training its machine learning models and justifying its retention practices.
The settlement proposal and complaint from the FTC highlighted the issue of companies amassing data, including from children, to refine their machine learning models. The FTC emphasized that compliance with data privacy laws is essential, and companies cannot use the excuse of improving algorithms to justify indefinite data retention.
Apart from the $25 million fine, Amazon will need to comply with several provisions outlined by the proposed order. These include prohibiting the use of certain data for the creation or improvement of data products, deleting inactive Alexa accounts of children, notifying users about the FTC-DOJ action, disclosing retention and deletion practices, and implementing a privacy program related to geolocation information.
Amazon stated that it disagrees with the FTC's claims but settled the matter to put it behind them. The company commits to removing child profiles that have been inactive for more than 18 months, although the FTC clarified that parents can choose shorter retention periods for their children's voice recordings. Amazon will continue to use the data to train its machine learning models, citing the purpose of providing and improving its services.
It's important to note that this settlement with the FTC is separate from another action involving Amazon subsidiary Ring, both of which highlight concerns about inadequate privacy and security protections.
Comments
Post a Comment