
Just as we have willfully given up a modicum of privacy in today’s age of social media, these devices are always listening. By design, they must be. Each manufacturer has a slightly different method and algorithm for what they do with that information. Most only record a short segment of what is being said, sending that to a cloud-based server and then overwrite that with the next 30 (or so) seconds of speech. However, what you say can and will be used against you.
Google, for instance, will pick up key words in the speech it hears and then send you targeted advertisements to your web surfing, particularly if you are using their browser Chrome. Echo will send you ads to your Facebook account if it is linked. Don’t be surprised if coupons for pizza show up in your feed if you were recently talking about it with one of these devices in the room. It happened to me last night after I mentioned in a phone conversation that we had pizza for dinner — blamo — Domino’s ad in my Gmail inbox. I have both an Echo and a Google Home device sitting next to each other for a research project, but I suspect it was the Google Home since the ad went to my Gmail.
It can also tell the difference in voices, I assume by speech pattern and frequency. If my 13 year old asks one of the devices to “play music,” the latest teen screamer comes on. If I ask I get a hearty round of blues classics. My wife won’t talk to either device. So how does this bode for privacy and security? We may soon find out.
A recent murder investigation in Arkansas has prompted the police investigating the crime to file a search warrant to prod Amazon into trying to recover the data from their cloud storage. Amazon has turned over basic information, but like in the recent Apple vs. FBI case, refuses to provide customer data with a specific court order. The slope is definitely slippery as who is to say that someone won’t demand that all speech be stored and be made available on demand. I am I not just referring to the government law officials.
A challenge I see is: What if the device picks up a voice from a TV where someone is watching the latest slasher flick? Would just text be available or would voice be part of the package? Additionally, some manufacturers of these types of devices outsource speech recognition and cannot actively account for the process and collection method. Samsung’s Smart TV with Voice Recognition was just dinged for this.
The technology is still in its infancy but is moving quickly. A look at Samsung’s recent purchase of Harman, which includes AMX, which happens to have a relationship with IBM and its Watson AI project, is a good example of where this may be headed. Their goal is big data collection to sell you more stuff. But with big data comes big responsibility. We just have to decide how much of our privacy is worth the convenience.