Hi, hmm, I’m a Robot

Google I/O was last week. I’ve only watched a tiny part and it’s probably the one you’re most likely to have seen as well: when the Google Assistant calls a haircut salon, then a restaurant, and successfully pretends to be a human.

It was mind-blowing. It also rushes us towards the future and poses some fascinating questions: should the Google Assistant identify itself when making these calls? Or is it OK to fool the human on the line? I’ve seen people call it a question of ethics, but I’m torn.

When a robot calls us with a robotic voice — and automated calls are nothing new — we’re fine with it not explicitly identifying its nature because it’s obvious. The difference here is the robot sounds perfectly human, which means 2 things can happen over the course of a conversation:

  1. A successful outcome is achieved with the human still blissfully unaware that they’ve been talking to a robot.
  2. Maybe it’s a few awkward moments, maybe the robot makes no sense, without being able to confirm it, the human isn’t sure they’re talking to a real person.

In case 1 I think it’s fine that the robot didn’t disclose anything. Case 2, however, is a real conundrum. Especially in the case where the person represents a business and is answering a potential customer: If the conversion becomes non-sensical and they know for a fact they’re talking to a robot, they can hang up right away. If they’re talking to a human, even if the conversion becomes weird, they would probably hang in there a bit longer and practice some patience and compassion.

What then? Perhaps the Google Assistant should identify itself when asked, “Are you a robot?” In fact, I suspect more and more humans will have to answer that question in the coming years. Perhaps the Assistant needs to be extremely good at identifying confusion and deadlocks so that they can exit the conversation gracefully.

The obvious option is for robots to be upfront about what they are. And yet something feels discriminatory about requiring an intelligence to identify itself as being artificial because we find them to be different or inferior.