Last week Google demoed Duplex, an AI intended to make phone calls on users’ behalf. Depending on who you ask it’s either mind blowing or dystopic.
Here’s the presentation, if you missed it—the demo starts around 1:12 if you’d like to skip the preamble.
The first thing you’ll notice: this thing is pretty good at thinking on its feet. This is a recording, not a live demo, meaning odds are this is a best case scenario for the tech. But it’s not hard to imagine something like this being useful, especially if you hate making phone calls. Read more on the Google Blog if you’re interested about how this all works, there are a couple more demos there too.
More than a few people noticed Duplex inserting filler words like “umm”, which is clearly intended to make the machine seem more human. Is that a useful feature, making the interaction more seamless, or a deliberate deception?
I am genuinely bothered and disturbed at how morally wrong it is for the Google Assistant voice to act like a human and deceive other humans on the other line of a phone call, using upspeek and other quirks of language. "Hi um, do you have anything available on uh May 3?" #io18
— Bridget Carey (@BridgetCarey) May 8, 2018
This stuff is really, really basic, but: any interaction with technology or the products of tech companies must be exist within a context of informed consent. Something like #GoogleDuplex fails this test, _by design_. That's an unfixable flaw.
— Anil Dash (@anildash) May 9, 2018
Google put out a statement later saying Duplex would identify itself as an AI, which satisfies some of these complaints. But there’s something unsettling about AI pretending to be human in vaguely social situations, and I think this kind of pushback is only going to get louder as the tech improves.