Duplex shows Google failing at ethical and creative AI design | TechCrunch

Yet despite all the thoughtful ethical guidance and research that’s already been produced, and is out there for the reading, here we are again being shown the same tired tech industry playbook applauding engineering capabilities in a shiny bubble, stripped of human context and societal consideration, and dangled in front of an uncritical audience to see how loud they’ll cheer.

Google’s Duplex demo was very impressive on a technical level. Even if it fails a bunch and the demos they showed were best-case scenarios (and they likely were), it is amazing to see a computer navigate conversations like the ones they showed on stage.

As we all know, there has been quite a bit of debate around this feature. Some say that adding “Hello, this is the Google Assistant” at the start of the conversation would solve the ethical problem here. I personally fall into that camp, as it’s not the function that’s distressing, it’s the deception.

But then I think about how people interact on the phone to computers vs humans. We’re more likely to be patient with humans and speak to them more naturally than computers. If people hear “this is Google Assistant” at the start of a call, are they going to be short with it? Will they prank it with bad information? Will they just hang up the phone because they don’t care for this one bit? Maybe this tech relies on people thinking they’re talking to another human being to work.

On the other hand, we all have a voice assistant in our phones and speakers and we seem to be able to interact with them relatively well¹, so maybe that’s not a huge issue. I know that Siri and Alexa are computers, but I can interact with them just fine to get the information I need to to get them to do thing for me.

But our current interactions with voice robots are all initiated by us. I’m talking to Siri because I want to talk to Siri. If Siri started to talk to me out of the blue I probably would be less enthused about talking to it. Businesses will have more incentive to be nice since they can make more money by cooperating with the bot and getting the task completed, but I still think there could be an attitude of “this is a computer, not a person, so I’m not going to take this call as seriously.”

This is a complicated topic, which is why it has generated so much conversation this week. A part of me is wildly impressed by this technology and am excited to see our voice assistants improve at a rapid rate. I want this stuff to get better and make our lives better. On the other hand, I am distressed by Google’s apparent “build it first, then apply ethics” pattern to be a bad look for the company and the industry as a whole.

https://twitter.com/mattbirchler/status/993908853648625664