BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Could Google's Duplex Rid Us Of Robocalls?

Following
This article is more than 4 years old.

Getty

Over the last several years robocalls have gone from minor nuisance to growing threat, not just to the public’s trust in Caller ID and anonymous calls but for their steadily increasing ability to impersonate the companies and government entities we do business with in order to conduct identity theft and other crimes. Could a technology similar to Google’s Duplex actually offer a powerful tool to fight back?

Google’s Duplex technology was announced last year as a way to enable the company’s digital assistant products interact more naturally with the real world, seeking out information on their own. The basic idea is that a user wanting to book a restaurant table or schedule a hair appointment could simply ask their assistant to do so and it would utilize Duplex to call the business and engage in a natural and convincingly human-like dialog with its human staff in order to complete the desired tasks.

What if a Duplex-like system could be applied in reverse? Instead of making calls to others, what if it answered all incoming calls like a receptionist and screened them to allow only relevant calls through?

Imagine Duplex answering every incoming call. Calls from known numbers for whom the caller’s voiceprint matches the typical voiceprint from that line would be passed through automatically, but all calls from unknown numbers would be answered by a digital receptionist that engages with the caller to both confirm it is a human calling and to take down their information.

Automated notification calls like appointment reminders could simply be passed through to the user’s calendar or provided as a transcribed summary text message.

All other robocalls, whether connected to a human or a machine, would result in a brief conversation by the digital receptionist that takes down the caller’s information and the purpose of their call. Certain kinds of calls could be blocked automatically through a user-defined blacklist. For example, a user could instruct the system to block all timeshare calls. Calls for other topics could result in a summary message displayed to the user on their phone screen that they could use to either dismiss the call or decide to accept it and be connected to the caller.

Such a system would incentivize callers to provide false information to the system in hopes of being connected to the human user, but as the conversational skills of such dialog systems improve, they could gradually learn to more robustly filter calls.

Of course, in the cat-and-mouse game of robocalls, it is almost a given that robocallers themselves will eventually make use of such technologies to scale their calling operations even further, entirely eliminating the need for human operators until the final moment of a sale. If consumers began adopting screening technology in large numbers, it is likely robocallers would respond in kind even faster.

Putting this all together, the FCC’s push towards automated behavioral screening of calls and the industry’s move towards more secure calling technologies will hopefully eventually curtail the most egregious robocalling practices, but the idea of using machines to screen our calls is an intriguing one.

Perhaps someday in the future our phones will be intelligent enough to hold conversations on our behalves, relegating junk phone calls to the spam folder of history.