What happens when a robot tries to sell you insurance?

What happens when a robot calls you up? It may sound like a bizarre question, but that’s exactly what writers from TIME Magazine asked themselves after Samantha West phoned them in 2013.

West claimed to be selling insurance, but something about her manner didn’t seem right, according to TIME staffers Zeke Miller and Denver Nicks. When asked point blank if she was a robot, West refused to give a straight answer.

After posting recordings of conversations with West, TIME’s staff got to the bottom of the mystery. The insurance saleswoman was indeed human. Or partly, at least.

Offshore workers, it turned out, were using pre-programmed voice recordings to provide human-like responses during sales calls. Case closed.

Only now, in 2017, there is an increasingly real prospect that a modern Samantha West could in fact be a robot.

Machine or human?

Today, says Wendell Wallach—author of A Dangerous Master and senior adviser to The Hastings Center—you might be hard pressed to tell if a Google translation had been carried out by a machine or simply by someone with poor language knowledge.

“That shows just how far we’ve come,” Wallach says. “It wouldn’t be hard to hook that up to a synthesizer.”

Wallach believes a smart human could still spot a robot easily but this may not be the case in the near future.

“Perhaps that space is closing more quickly than we might imagine,” he says.

Sebastian Reeve, director of product management at Nuance Communications, which makes voice systems for customer service, confirms that today’s robots are a lot more sophisticated than they were in 2013.

“[Virtual assistants] are constantly learning from their experience,” Reeves says. “There’s no question that these bots can very quickly become self-reliant in responding to a great number of questions and queries.”

The trend poses important questions for technologists and policymakers. Wallach, for example, feels it may soon be time to introduce regulation that obliges bots to state what they are—as TIME reporters asked Samantha West to do in 2013.

Increasingly on guard

In the meantime, the advent of human-robot communications means consumers must increasingly be on their guard.

And with many financial transactions now taking place via telephone or online channels, it is more important than ever for customers to deal only with trusted service providers. That won’t necessarily stop you from being contacted by a robot, of course.

So, if you’re mid-conversation with someone who sounds a bit like Samantha West, how can you tell if they are made of flesh and blood?

“The problem is, ‘You’re a robot!’ is pretty damn insulting,” concedes Wallach. “There’s an area of courtesy there.”

Instead, he says: “I would try to do something that put us in a totally different conversational style, so we’re talking about a totally different topic.”

Interrupting the caller to say you have another call coming in from the hospital, for example, might throw a bot’s circuits off guard in a way that betrays their identity.

Just don’t ask if they are a robot.

Share this article:
Share on LinkedInTweet about this on Twitter