Google data privacy suit alleging violations from voice AI training partially dismissed
A data privacy lawsuit against Google over the speech recognition processes used by its voice assistant services has been partially thrown out by a federal judge in California, but the suit will be allowed to go forward, and could be amended.
U.S. District Court Judge Beth Labson Freeman ruled that claims Google violated federal privacy law by disclosing user conversations to third-party contractors, MediaPost reports, but claims made under federal wiretap laws and California’s Consumer Privacy Act have been dismissed, according to Reuters.
The contractors received voice recordings from Google to analyze language patterns, specifically for improving the performance of its speech recognition with different languages. The suit was filed shortly after the company’s use of third parties and user data for training its voice services was reported in Europe last year. Some of the conversations were also recorded without consumers using the hotwords that initiates voluntary interaction with the device.
Google argued in federal court that the case should be dismissed entirely due to two provisions in its privacy policy, in which the company sets out the possibility of sharing user data with outside contractors and using data to improve its services. The judge ruled that while users may have consented to the data collection for service improvement purposes, the consent cannot reasonably be extended to disclosure of the data, because that provision is included in a different section of the privacy policy.
Several other claims in the complaint were dismissed, but without prejudice, so they could be changed and refiled up until June 5.
Google remains embroiled in lawsuits alleging violations of Illinois biometric data privacy law by both its photo service and its classroom tools, with the latter suit also alleging federal law violations.
Article Topics
biometric data | biometrics | data protection | dataset | Google | lawsuit | privacy | speech recognition | training | voice biometrics
Comments