Mobile Conference - A revolution in the user experience

The Samsung Smart TV Privacy Issue Represents a “Deep Misunderstanding of Reality,” According to William Meisel, Executive Director, Applied Voice Input Output Society

The Applied Voice Input Output Society (AVIOS), a 34-year-old industry organization, reacted to the spate of news that compared the voice control features of Samsung Smart TVs to Orwellian surveillance of everything we say, originally posted on Twitter by Parker Higgins, a member of San Francisco activist group Electronic Frontier Foundation. A privacy policy from Samsung warned users of a smart TV that the voice commands were handled by a “third-party” (later revealed to be Nuance Communications) and could pose privacy issues. Samsung later clarified that there were two microphones involved, one in the TV that allowed minimal control actions speaking across the room (where the voice signal is interpreted locally and never leaves the TV) and one in the remote control activated by a button press to start listening.

William Meisel, Executive Director, AVIOS, and author of the 2013 book, The Software Society: Cultural and Economic Impact (which discussed in part the tightening connection between humans and computers using human language), explained, “The implication in the original reporting that everything you say near the TV is sent over the web is erroneous. It’s no different than when you speak to Apple’s Siri, Microsoft’s Cortana, or Google’s voice search (or similar “personal assistants”) on a mobile phone. You activate the ‘listening,’ and the speech is sent over the Internet to be processed and to provide a response. If you dictate a text message to Siri or launch a Web search through Google’s voice search, the content is only private to the extent that Apple or Google protects that privacy. And it’s no different than a normal typed text message or typed search request, which is only private to the extent that the company processing the text keeps it private.” Meisel continued, “The implication that voice is somehow different than the general issue of privacy in our digital systems is simply wrong. To the degree that voice is more difficult to review in volume than text, one could even argue that voice is more private than text.”

Part of the reason for Samsung’s disclosure is that they are using the speech recognition technology of Nuance Communications (a “third party”) to perform the speech recognition and content lookup in the cloud, similar to the way that Google processes voice in the cloud when you use voice search. Samsung is revealing that the voice signal leaves their corporate control, and that, in effect, it is Nuance protecting the privacy of the voice commands rather than Samsung. K. W. 'Bill' Scholz, president, AVIOS, and president, NewSpeech, noted that today, this “third-party” handling of data is a widespread trend. “The growth of cloud services, where another party handles the data of a company during processing of that data, is today more the practice than the exception. For example, a number of companies provide basic data processing in the cloud as well as handling specific types of inquiries in web interactions. Data is data, and privacy is an issue that all companies must deal with, whether that data is text or voice and whether they handle that data only within the company or get help from a third party.”

Much of the press coverage represents a deep misunderstanding of reality, AVIOS claimed. Meisel commented, “Personally, I don’t think there is anything that is revealed in the voice commands I give my TV other than perhaps my bad taste in entertainment choices.”

About the Applied Voice Input Output Society

The not-for-profit Applied Voice Input Output Society is an industry organization that has promoted the development of commercial applications of speech and language technology for 34 years. The organization’s fifth annual Mobile Voice Conference will be held April 20-21, 2015, in San Jose, California. See