Amazon has been given a new patent, which would permit Alexa to identify many different user features, such as emotional state and accent. According to The Intercept, it could give rise to related facial recognition tech civil rights problems, including apprehensions about privacy, surveillance, and discriminatory advertising.
The abilities mentioned in the patent include recognizing consumer’s accent, ethnicity, emotional state, age, gender, and background noise.
Data and privacy protection rules were not made by keeping in mind such abilities, and Andrew Crocker, Senior Staff Attorney of Electronic Frontier Foundation (EFF), claims that the Foreign Intelligence Surveillance Act (FISA) provides authority to the US government to secretly ask for data access, like accent or ethnicity, which may assist recognize possible terrorists. Privacy scholars also observe the potential for the ability to be utilized for control implementation and immigration.
Amazon also does not tell the full info it receives about National Security Letters (NSLs). The Intercept claims that surveys from Facebook, Apple, and Google propose that the NSLs’ use to compel info from tech firms is rising. Amazon information request surveys point to a prominent rise in the number of orders it got for consumer data in 2015, which is the same year it releases the Echo intelligent home speaker.
The Electronic Communications Privacy Act (EPCA) that was approved in 1986 supplies few securities for voice data gathered by Amazon, but they are restricted, and according to The Intercept, privacy scholars slightly differ on their analysis of the extent of these securities. Recently, a new judge in Hampshire asked Amazon to give the recordings from Echo for investigating a murder.
Human Rights Watch report has been cited by The Intercept, which says that the Chinese govt. has claimed to have solved many criminal cases by utilizing tech from voice biometrics provider iFlyTek. Another cited report, Die Welt, says that the German govt. has used voice analysis software on a trial basis to recognize the hometown of refugees.
Tractica, an analyst firm, has forecasted the worldwide market for speech and voice recognition tech to be of $6.9 billion in value by 2025.