A recent advertising campaign from Samsung promises: “TV has never been this smart.” We may soon wonder how smart a TV can get before it is too smart for our own good.
“Smart TVs” can display Web content by directly accessing a home’s Internet connection. Many of the high-end versions, including Samsung’s, offer voice-recognition technology, allowing users to change channels, search for programs or adjust the TV volume by verbal command.
Samsung is not the first company to introduce voice controls for smart TVs, but it is the focus of a privacy group’s current concern. The Electronic Privacy Information Center has asked the Federal Trade Commission to investigate after a close reading of the company’s privacy policy revealed that third parties might be able to listen in using the TV’s built-in microphone. The policy warns users 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.”
While we probably don’t have to worry about Skynet rising in our living rooms, Americans are rightfully wary of sweeping electronic intrusions. After the revelations about National Security Agency programs and privacy breakdowns at companies as diverse as Anthem and Target, it is understandable if some consumers are unsettled by the idea of an unnamed person listening in on conversations that happen to take place while the TV is on. And smart TVs are not the only culprit: Some video game consoles can be controlled by voice too, and Apple and Android mobile devices have made voice search an everyday part of many people’s lives.
Nearly all of these devices let you turn the microphone off or disable voice recognition software. You can also disconnect your smart TV from the network so it won’t transmit anything, though this means losing most of the benefits of owning a smart TV. And most devices that use voice controls require an initiating command before they start recording what you say at all, at least in theory.
For TV owners who choose not to disable voice controls, the Samsung policy still may not be cause for undue alarm. A Samsung spokeswoman, speaking to Chris Matyszczyk at CNET, explained that the third parties the policy indicates are contacted only during a requested voice command search; no voice data is retained or sold, she stressed. But while this is reassuring, it requires customers to trust the company not to retain or sell collected data in the future.
TV buyers may be wary, especially if they are aware of the investigation LG triggered a few years ago when it came to light that viewing data from LG’s smart TVs was collected even if the related setting was toggled to “off.” (LG later released an update to fix the issue.) There has also been concern that Samsung’s data, when transmitted, is not properly encrypted. Companies need not be malicious to compromise consumer privacy; they need only be sloppy.
Apple and Google have been careful to specify that data from smartphone voice search or services like Siri is anonymized, so the companies cannot trace a given query back to any particular user – in Google’s case, ever, and in Apple’s case, after six months connected to a randomly generated number. But even without personal information attached, sensitive data sitting on a company’s servers could be a problem. For instance, a dictation might contain legally regulated information, such as the precise time a company plans to file for an initial public offering. Scrubbing the name of the person who input the data may not be enough when the data itself needs protection.
There are a few ways the privacy concerns might eventually be resolved. A device maker may find itself held financially responsible if it obtains a certain sort of sensitive information - for instance, information about planned criminal activity - and fails to take responsible action. If and when this happens, that manufacturer is likely to promptly disable or eliminate voice data gathering capabilities. Wary competitors would likely follow suit rather than risk ending up in the same legal hot water.
It may not come to that, of course. Device makers may voluntarily limit where our data goes, or legislators may force them to do so. After all, gathering information as such is not the problem. Obviously, we realize that when we ask our smartphones for directions and traffic information, the phone must communicate the request to an outside server; voice recognition simply acts as a fingerless keyboard for inputting search queries. The phone itself doesn’t “know” the answer. It relays your request to an app or a search engine and returns the answer to you.
Similarly, if I ask my smart TV to display a channel guide or play a particular program, I know it is obtaining the content from elsewhere. That is not a problem. In fact, it is probably why I purchased a smart TV in the first place. The problem is that consumers are not necessarily agreeing to let the TV maker store that data, ostensibly for product-improvement purposes, or share that data with third parties for marketing purposes.
If I search for a certain website on my MacBook, I don’t expect Apple to be informed. I have no reason to expect this on my iPhone either, whether I use voice technology or my fingers to enter the site name. Companies need research to improve products, of course - but they can conduct this research in house, or use beta testers who know their usage is being monitored. There is no reason to turn the entire customer base into unpaid research assistants, even though such practices are now common.
The solution may eventually come from technological progress itself. One day, machines may have the storage and processing capacity to handle all voice commands locally, eliminating any need to transmit the spoken commands (or transcriptions of spoken commands) elsewhere. The more that can be wired into the hardware, the less need to move data or to involve third parties.
In the meantime, devices’ programming should limit transmitted information to phrases that are recognized as some part of the unit’s functionality. There is no need to record or transmit phrases such as “my husband is a pompous idiot.” Our gizmos should be smart enough to tell the difference.
Larry M. Elkin is the founder and president of Palisades Hudson, and is based out of Palisades Hudson’s Fort Lauderdale, Florida headquarters. He wrote several of the chapters in the firm’s recently updated book,
The High Achiever’s Guide To Wealth. His contributions include Chapter 1, “Anyone Can Achieve Wealth,” and Chapter 19, “Assisting Aging Parents.” Larry was also among the authors of the firm’s previous book
Looking Ahead: Life, Family, Wealth and Business After 55.
Posted by Larry M. Elkin, CPA, CFP®
photo by Karlis Dambrans
A recent advertising campaign from Samsung promises: “TV has never been this smart.” We may soon wonder how smart a TV can get before it is too smart for our own good.
“Smart TVs” can display Web content by directly accessing a home’s Internet connection. Many of the high-end versions, including Samsung’s, offer voice-recognition technology, allowing users to change channels, search for programs or adjust the TV volume by verbal command.
Samsung is not the first company to introduce voice controls for smart TVs, but it is the focus of a privacy group’s current concern. The Electronic Privacy Information Center has asked the Federal Trade Commission to investigate after a close reading of the company’s privacy policy revealed that third parties might be able to listen in using the TV’s built-in microphone. The policy warns users 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.”
While we probably don’t have to worry about Skynet rising in our living rooms, Americans are rightfully wary of sweeping electronic intrusions. After the revelations about National Security Agency programs and privacy breakdowns at companies as diverse as Anthem and Target, it is understandable if some consumers are unsettled by the idea of an unnamed person listening in on conversations that happen to take place while the TV is on. And smart TVs are not the only culprit: Some video game consoles can be controlled by voice too, and Apple and Android mobile devices have made voice search an everyday part of many people’s lives.
Nearly all of these devices let you turn the microphone off or disable voice recognition software. You can also disconnect your smart TV from the network so it won’t transmit anything, though this means losing most of the benefits of owning a smart TV. And most devices that use voice controls require an initiating command before they start recording what you say at all, at least in theory.
For TV owners who choose not to disable voice controls, the Samsung policy still may not be cause for undue alarm. A Samsung spokeswoman, speaking to Chris Matyszczyk at CNET, explained that the third parties the policy indicates are contacted only during a requested voice command search; no voice data is retained or sold, she stressed. But while this is reassuring, it requires customers to trust the company not to retain or sell collected data in the future.
TV buyers may be wary, especially if they are aware of the investigation LG triggered a few years ago when it came to light that viewing data from LG’s smart TVs was collected even if the related setting was toggled to “off.” (LG later released an update to fix the issue.) There has also been concern that Samsung’s data, when transmitted, is not properly encrypted. Companies need not be malicious to compromise consumer privacy; they need only be sloppy.
Apple and Google have been careful to specify that data from smartphone voice search or services like Siri is anonymized, so the companies cannot trace a given query back to any particular user – in Google’s case, ever, and in Apple’s case, after six months connected to a randomly generated number. But even without personal information attached, sensitive data sitting on a company’s servers could be a problem. For instance, a dictation might contain legally regulated information, such as the precise time a company plans to file for an initial public offering. Scrubbing the name of the person who input the data may not be enough when the data itself needs protection.
There are a few ways the privacy concerns might eventually be resolved. A device maker may find itself held financially responsible if it obtains a certain sort of sensitive information - for instance, information about planned criminal activity - and fails to take responsible action. If and when this happens, that manufacturer is likely to promptly disable or eliminate voice data gathering capabilities. Wary competitors would likely follow suit rather than risk ending up in the same legal hot water.
It may not come to that, of course. Device makers may voluntarily limit where our data goes, or legislators may force them to do so. After all, gathering information as such is not the problem. Obviously, we realize that when we ask our smartphones for directions and traffic information, the phone must communicate the request to an outside server; voice recognition simply acts as a fingerless keyboard for inputting search queries. The phone itself doesn’t “know” the answer. It relays your request to an app or a search engine and returns the answer to you.
Similarly, if I ask my smart TV to display a channel guide or play a particular program, I know it is obtaining the content from elsewhere. That is not a problem. In fact, it is probably why I purchased a smart TV in the first place. The problem is that consumers are not necessarily agreeing to let the TV maker store that data, ostensibly for product-improvement purposes, or share that data with third parties for marketing purposes.
If I search for a certain website on my MacBook, I don’t expect Apple to be informed. I have no reason to expect this on my iPhone either, whether I use voice technology or my fingers to enter the site name. Companies need research to improve products, of course - but they can conduct this research in house, or use beta testers who know their usage is being monitored. There is no reason to turn the entire customer base into unpaid research assistants, even though such practices are now common.
The solution may eventually come from technological progress itself. One day, machines may have the storage and processing capacity to handle all voice commands locally, eliminating any need to transmit the spoken commands (or transcriptions of spoken commands) elsewhere. The more that can be wired into the hardware, the less need to move data or to involve third parties.
In the meantime, devices’ programming should limit transmitted information to phrases that are recognized as some part of the unit’s functionality. There is no need to record or transmit phrases such as “my husband is a pompous idiot.” Our gizmos should be smart enough to tell the difference.
Related posts:
The views expressed in this post are solely those of the author. We welcome additional perspectives in our comments section as long as they are on topic, civil in tone and signed with the writer's full name. All comments will be reviewed by our moderator prior to publication.