Seldom does a paragraph in a TV’s User Manual go viral. Yet that’s exactly what happened when the Electronic Frontier Foundation compared the following line from Samsung’s “privacy policy” to Orwell’s description of “telescreens” in the novel 1984:

“Please be aware that if your spoken words include personal or other sensitive information, that information will be among the data captured and transmitted to a third party through your use of Voice Recognition.”

This disclosure only fulfills half of the requirement for “informed consent,” in that it lets TV owners know that their new appliance is listening to their conversations, but it doesn’t really provide any recourse or controls over how to prevent it. Indeed, it took Samsung less than a day to take to the blogosphere to explain its technology and re-write its privacy policy.

Under the headline “Samsung Smart TVs Do Not Monitor Living Room Conversations” and the byline samsungtomorrow, company spokespeople explain that its SmartTVs perform two very different types of automated voice recognition. One is of the embedded variety and supports a limited vocabulary of commands and phrases. The other, which is supported by a microphone in the remote control and can be activated or de-activated at any time, transmits utterances into “the cloud” (in this case Nuance Communications’ remote resource for speech-to-text transcription).

As I’ve noted in our ongoing coverage of “The Conversational Cloud” this is where the magic happens in terms of Intelligent Assistance (IA). Resources for transcription of spoken words reside in close proximity to the servers that perform natural language processing, machine learning, conversation management and speech analytics. Or, as Samsung explains, it lets one of the TV owners say “Recommend a good Sci-Fi movie” and get results on screen.

Critics are up-in-arms about the invasion of privacy entailed in automatically uploading spoken words to third-party resources. TechCrunch’s Natasha Lomas, for one, makes an articulate case for baking “Privacy by Design” into all consumer electronics, automotive and mobile resources that rely on sending personal information into third-party or public clouds.

Yet there are even deeper issues exposed by the Samsung privacy debacle. It probably isn’t a coincidence that the viral privacy policy coincided with a segment of CBS’s 60 Minutes that showed how a DARPA researcher (and presumably the NSA or a malicious hacker) could use communications links from the cloud to take control of your automobile. The implications are truly frightening and would give anyone pause before making the choice to by a “connected car”, “smartTV” or any other platform for intelligent assistance.

Tom Goodwin of Havas Media had it right when he published an article in Advertising Age called, “Intimate Data Will Be Key to the Internet of Things.” In it, he notes that advertisers need to be more people centric, noting that advertisers “need to work around people and new calls to action.” To me, it is a scary thought to think that advertisers might be a few steps ahead of us in taking control of our own data. In that respect, they are no different from malicious hackers.

Here’s my call to action: For companies that are launching “smart” devices on the Internet, be transparent about why, when and how you collect and use personal information. For individuals putting these devices to use, demand to know how you can control when and what you are revealing about yourself. To both, strike a balance between the costs (in terms of loss of privacy, investment in time, exposure to hacking, etc) and the benefits of using devices and services that involve selective disclosure of personal information. It’s a constant trade-off, but it is important for individuals to understand that there are real advantages to engaging with their own virtual intelligent assistant or assistants.

The challenge is not about technology, privacy or compliance. It’s about people. Intelligent Assistance need to be of real value to individuals. And service providers have to explain how it saves time, money and generally improves lives. And then they have to deliver on their promises. Overcoming objections surrounding privacy and malicious hacking is important but, if IAs prove their worth (which I’m confident they will) those objections will be relegated to background noise.

Categories: Intelligent Assistants

Tags: Intelligent Assistance, Voice Recognition, Samsung