Amazon is sued for storing unauthorized child recordings

Two lawsuits have been filed against Amazon for storing children’s recordings through its smart speaker, report experts in personal data protection. Claimants allege that the company does not have the consent of users to create voice records with which it could track the activity of an infant in the services enabled for voice assistant Alexa, which would allow the company to elaborate detailed profiles about children.

The lawsuits, presented in Los Angeles and Seattle, claim that two children used Alexia for Internet searching, play music and videos, among other interactions; the plaintiffs claim that they never gave the company consent to record a profile of the children. 

For its part, the company claims that it only stores the information when the user explicitly expresses their consent, in addition to the parents being able to delete any voice record that Alexa has stored.

According to the experts in personal data protection, the plaintiffs seek compensation for damages and ask interested users to join the collective lawsuit against Amazon for breach of privacy laws. Amazon mentions that, at the end of 2018, more than 100 million of devices had been sold with the integrated Alexa voice assistant, such as smart speakers, hearing aids, Smart TVs, among others.

The voice assistant software starts working once it hears its ‘wake word’ (e.g. Alexa, Amazon, computer, etc.). When the wake word is detected, the system starts transmitting the user’s words to the Amazon servers for processing their requests. However, these devices may fail at times, so many user voice records are transmitted to Amazon without the user having said the activation word.

According to the experts in personal data protection from the International Institute of Cyber Security (IICD) users of this system can prevent this from happening by denying consent; In addition, this service has an option that allows the user to enter a detailed record of his voice, repeating several sentences for the wizard to detect the voice of the user with a lower margin of error, which strengthens the Alexa machine learning process.