You probably already know the Internet is spying on you. Between the personalized ads, website cookies, and Alexas, we’ve signed over more than we know to companies looking to capitalize on our data. But now, Amazon is looking to literally listen to its consumers. The tech company has developed an improved “sniffer” algorithm, highlighting keywords in customers’ conversations to determine what the buyer wants and when.
While this may be shocking to some, Amazon has been listening to us for ages — now it just wants to be more precise about what it can glean from us. The company recently released its Halo fitness tracker, which listens to your voice to gauge your mood and "scans" your body using photos to measure body fat data. Additionally, Amazon launched Care Hub, a remote caretaking experience that lets caretakers see a rundown of their loved one’s activity. As long as the spying is with good intentions and there's explicit consent, it’s okay, right? Right?! Hmm.
Amazon wants to hear from you — The new technology named “Keyword determinations from conversational data” is currently pending a patent with the U.S. Patent and Trademark Office. It’s not necessarily a new technology, but rather a continuation of similar “sniffer” algorithms Amazon has filed in the past. These algorithms “sniff” out topics a speaker is potentially interested in, and according to Amazon, also “attempt to identify trigger words in the voice content, which can indicate a level of interest of the user.” So if you “love” something whether than “like” it, the algorithm will be able to tell. Amazon’s patent application further states “a keyword that is repeated multiple times in a conversation might be assigned a higher priority than other keywords, tagged with a priority tag.”
Why is Amazon doing this? Well, the company complains that existing sniffer algorithms aren’t precise enough, and it wants “context” regarding users’ data. Right now, the only data Amazon has access to — that it’ll admit to at least — is what users specifically search for on its site, and what they buy. But they want to know why you’re buying mini erasers in bulk, whether they’re a gift or not, when next you might want to order them, and anything else that can help it sell more products.
The same patent application also inserts the possibility of “facial recognition, or another such process, [which] can be used to identify a source of a particular portion of audio content,” particularly “if multiple users or persons are able to be identified as sources of audio.” This way, Amazon can market the correct products to the interested party.
Straight out of Black Mirror — This algorithm can be used on an array of devices, not just Amazon-owned ones. Any computing device, be it a laptop, smartphone, or iPad, can actively listen to audio data for a user. In its patent application, Amazon provides a handful of examples explaining how the algorithm works – including ones that occur in real-time.
One such example is if, hypothetically, a user mentioned a restaurant while on the phone with a friend. If the user mentions a certain restaurant, Amazon might “send a recommendation while the user is still engaged in the conversation that enables the user to make a reservation at the restaurant, or provides a coupon or dining offer for that restaurant (or a related restaurant) during the call.” The marketing would happen in real-time as after the call might be too late for the customer to make plans.
Hopefully the technology will have to clear regulatory hurdles before it's rolled out and Amazon will have to get explicit consent from users. An Amazon spokesperson told Input, “Like many companies, we file a number of forward-looking patent applications that explore the full possibilities of new technology. Patents take multiple years to receive and do not necessarily reflect current developments to products and services."
We hate it, though, nonetheless. And if it rolls out, we might think twice about putting Echo devices in our homes... or using Amazon's services in anything other than a soundproof room.