Slashdot Mirror
Google Is Being Vague With Disclosure In Early Real-World Duplex Calls (theverge.com)
A small group of Pixel owners in "select" U.S. cities are able to use Google's new Duplex feature to automatically make voice calls to restaurants and other businesses on their behalf. Referencing a demo from VentureBeat, The Verge notes that "the exchange between Duplex and a restaurant on the other side of the call is raising some early concerns about transparency." From the report: [Y]ou'll notice that Duplex never identifies itself as a robot. It never tells the person taking the call that they're interacting with an automated system. "Hi, I'm calling to make a reservation for a client. I'm calling from Google, so the call may be recorded," is what Duplex says to begin the conversation. And that little bit -- about the call coming "from Google" and potentially being recorded -- is the only disclosure that it ever provides. From then on, Duplex handles the requested dinner reservation smoothly.
This disclosure doesn't match up with a promotional video for Duplex that Google posted to YouTube back in June. In that example (embedded below), Duplex makes it very clear that it's a bot. "Hi, I'm the Google Assistant calling to make a reservation for a client. This automated call will be recorded." That's a much better approach. You're talking to the Google Assistant. It's an automated call, and it is being recorded; no maybes about it. The report notes that some Duplex calls -- such as the one VentureBeat included in their demo -- are actually handled by a human. "When a human operator at Google places a Duplex call, they don't necessarily disclose anything about Google Assistant or note it's an automated call," reports The Verge. "Because it's not. Not entirely, anyway. Google's Duplex tests involve a mix of the two; some are led by Googlers, while others let the AI steer. The majority of calls are the latter and automated, from what I'm told."
This disclosure doesn't match up with a promotional video for Duplex that Google posted to YouTube back in June. In that example (embedded below), Duplex makes it very clear that it's a bot. "Hi, I'm the Google Assistant calling to make a reservation for a client. This automated call will be recorded." That's a much better approach. You're talking to the Google Assistant. It's an automated call, and it is being recorded; no maybes about it. The report notes that some Duplex calls -- such as the one VentureBeat included in their demo -- are actually handled by a human. "When a human operator at Google places a Duplex call, they don't necessarily disclose anything about Google Assistant or note it's an automated call," reports The Verge. "Because it's not. Not entirely, anyway. Google's Duplex tests involve a mix of the two; some are led by Googlers, while others let the AI steer. The majority of calls are the latter and automated, from what I'm told."
If you like spam (the unsolicited email, not the spiced ham), go ahead and allow callers to shift the cost of handling calls entirely to you. They will not hesitate to have you called for minimal gain. This will kill the phone.
why do you have to know if you're talking to an AI or not?
i don't see how that even matters, i know people get worked up about it, but i don't know why.
On a long enough timeline, the survival rate for everyone drops to zero.
In a business setting, time taken to answer calls costs money. Personal calls just cost your time (just like spam email doesn't cost you anything but your time). It takes the callee time to answer the call, but unlike in a person-to-person call, it no longer takes the time of the "caller". This encourages the "caller" to have more calls made, which costs the callee more time and/or money. It's exactly the email spam dynamic.