Google stunned audiences last week as its remarkably human-sounding artificial intelligence software, Duplex, held actual conversations with presumably unknowing humans at a hair salon and a restaurant to make an appointment. Then, after the initial wave of amazement, the backlash began as people worried about the ethical implications of a human unwittingly being duped into interacting with Google’s virtual assistant.
I’m a firm believer that two things can be true at the same time. Yes, there are definite concerns surrounding the widespread deployment of such technology (the artificial “umms” to dupe the listener seem a little over the top), and I’m having a harder time thinking that the sky is falling.
Part of this stems from the fact that Google Duplex will be undoubtedly be used by restaurants as well. The software is still in development, and the only demo was around it being a (virtual) personal assistant for the customer. But Google is just a big data gobbler, and they aren’t going to only participate in half of this interaction, especially if there’s money to be made (and data to be had) in the other half.
As Jenn Marston wrote last week, almost 90 percent of consumers want to interact with businesses by text — sidestepping voice interaction altogether. And services like Guestfriend allow any restaurant to easily create their own chatbot to answer the same basic questions most customers ask, so automated assistance is going to become more mainstream for restaurant owners.
I assume that at some point, my virtual assistant will chat with the restaurant’s virtual host in some Matrix-like setting that may or may not end up in a gun-fueled kung-fu battle to get my 7:00 p.m. table.
On our internal Slack channel, The Spoon publisher Mike Wolf chimed in on this potential virtual assistant arms race, writing:
Today it’s pretty easy to get a seat at a restaurant, or even a decent restaurant. But what if people used AI-bots to do a high volume of calls/searches to make sure they have the best seat at the best time at the best restaurants? Or to get tickets for the local concert? Are we essentially arming people to remove any inefficiencies in a world? And does that leave those who don’t have access or the knowhow to employ a taskbot to pick up the scraps or eating at worse times or at worse places or getting whatever leftover appts there are at the styllist?
My point is: eventually the efficiency of automated taskbots will force others to use taskbots. The supplier of the service could create an impediment to this by requiring a “verified human” is making the actual engagement.
To get a more AI-focused perspective, I reached out to former colleague, Derrick Harris, AI expert and creator of the Architecht newsletter. He emailed the following, “My concerns probably mirror the ethical ones that other people have about forcing low-wage workers to try communicating with bots because rich people are too lazy/important to call.” He went on, “as we live more of our lives shopping online and not even making our own phone calls, do we look at people in the service industry as just things there to perform a service as part of digital transaction?”
But Derrick, Mike and I are all technology writers, so I talked with two people in the restaurant industry (a manager and a more corporate person) and asked them how they’d feel if they had to interact with a virtual AI-assistant. I know this is a very small sample size, and they didn’t want to go on the record, but their responses were illuminating.
The manager would be fine interacting with a virtual assistant who called in, as long as it didn’t take more time than talking to a human. If the AI couldn’t understand responses, or took to long, that would be frustrating and probably lead to the manager hanging up. They didn’t think that the assistant would be able to handle the more complex aspects of taking a reservation, like handling questions about seat preference and food allergies. (From a technology perspective, this doesn’t seem that hard to overcome.)
The more corporate person was from a different country. While their English was fluent, there is a noticeable accent, which, they said has caused audible frustration with people on the other end of the phone. The idea of having a virtual entity that could be “understood” more easily for basic tasks like booking a table was exciting.
The issue of AI in our daily lives isn’t going away, and it’s something we’ll have to grapple with as it improves. But it’s one thing for me, as a guy who only writes about food and technology to wax on about it, I’m more interested in hearing from the folks who will have to deal with it. With that in mind, if you’re in the restaurant industry, I’d love to know what you think about the idea of interacting with AI assistants rather than people. Do you hate the idea? Love it? Don’t care? Leave us a comment below!