Do Androids Dream of Cocktails? DnA Meets IBM’s Watson

Written by
Watson went head-to-head with taco chefs at IBM Cognitive, suggesting additional ingredients like fried olives and chopped nuts. (The original image is no longer available, please contact KCRW if you need access to the original image.)

You may have read about the South Korean Go champion who just lost several games to Google’s Alpha Go, prompting one South Korean journalist to write “Koreans are afraid that artificial intelligence will destroy human history and human culture.”

Perhaps that brings back memories of Watson, IBM’s supercomputer that beat out some highly educated humans at Jeopardy a few years back.

Well, Watson was one of the attractions at IBM Cognitive, an installation at South By Southwest Interactive, which just wrapped up in Austin, Texas, and  DnA got to test his powers of assisting with useful services, like mixing a cocktail and choosing a gift.

But did Watson get it right, and do we need help from the machine?

I was at SXSW to host a panel at Dell computers about the the Future of Making in the “Third Wave” of computing (post-screen, immersive digital engagement; more on that later) and while wandering in Austin’s downtown I saw a bunch of people lined up outside a bar temporarily renamed IBM Cognitive Studio.

They were being interviewed by staffers who punched responses into tablets so I joined the line and found myself being asked a bunch of questions about how I felt today, and was asked for my Twitter handle. Then I was given a wristband and lead to the bar where I had to swipe my wristband and then a bartender named Alex glanced at a screen and then produced a cocktail.

Visitors to IBM Cognitive Studio have to answer questions before entering. Their “personality profile” is crunched by Watson into suggested ingredients for a cocktail. (The original image is no longer available, please contact KCRW if you need access to the original image.)

Turns out the ingredients — gin, mint, grapefruit and jalapeno — were recommended by Watson, based on crunching my twitter feeds and the info about my state of mind that day into a “personality profile,” in which I was both bubbly and spicy.

The drink was tasty but I couldn’t help wondering why Alex, who seemed like a nice guy to chat with at a bar, and already knew how to make a mean cocktail, needed help from some artificial intelligence. The same applied to the chefs on hand who got helpful advice from Watson on what ingredients to mix into their tacos in a human-V-Watson cook-off.

Apparently Watson thought fried olives would be a big improvement on regular ones in one taco and that another could use some chopped nuts.

Cameron O'Connor with Peppa
IBM Program Director Cameron O’Connor stands with Pepper, the adorable face of AI, produced by IBM in alliance with SoftBank Robotics Holdings (SBRH). (The original image is no longer available, please contact KCRW if you need access to the original image.)

According to Cameron O’Connor, the very personable Program Director for IBM’s World Conferences, “the whole deal and the whole experience that we want to get across — especially with Watson — is about augmenting decisions, it’s not about making decisions for humans. It’s about giving humans an opportunity to see what other options might be out there.”

Those augmented decisions, he says, apply to lifesaving applications such as healthcare; IBM is working with oncologists at Sloan-Kettering to apply Watson AI to personalizing cancer treatments.

And it applies to retail. My next encounter with Watson was at a “Makers Mosaic,” (see Cameron O’Connor explain it in the video below) where I tapped out responses to questions about my 11-year old daughter’s personality.

Visitors at the Makers Mosaic at IBM Cognitive
Visitors at the Makers Mosaic at IBM Cognitive (The original image is no longer available, please contact KCRW if you need access to the original image.)

I had to answer things like, is she easygoing or is she organized? Is she reserved or outgoing? Confident or sensitive? Watson crunched the responses and up popped a black sweater with a huge yellow-gold heart on the front; price tag $59.

I was sure that my daughter would have been horrified by this sweater. To prove it, when I got back to LA I found a similar one on line, albeit with a white heart, and asked her what she thought of it.

To my astonishment she said she liked it, adding “well, it’s not my usual style but I like it.”  Then I asked if she would like it if the heart was yellow. “No!” Maybe Watson was half-right.

So, is this augmented decision-making creepy or helpful?   You may be able to find out for yourself soon enough, at a retail outlet near you. A pilot was tested this past Christmas holiday at a Westfield shopping mall.

As for Alex the barman, he says he imagines some bars will introduce the technology in coming years. But what are its advantages for him? He said, “it saves time. You ring up and basically it can tell me your flavor profiles and what you like but it’s also going to base it on on the mood you’re in aswell and the computer will draw me up a list of ingredients and then I get free reign to work on those ingredients.”

I guess that’s a benefit though surely one of the attractions of going to a bar is for the social connection, with friends and with bartenders? Do we want mixologists to mix our drink while staring at a screen for suggested ingredients?

IBM is aware that what people seek in life, and therefore in AI, is empathy and warmth, so Watson sometimes takes the form of a robot. One on hand was called Pepper and she/he is very cute and designed to help out on retail floors, assisting customers with their needs. Given the already minimal staffs in many stores, I couldn’t help wondering if helpful, and cheaper, Pepper might be replacing humans in these service jobs.

I didn’t get to test Pepper’s abilities, however, because at the moment I arrived, she/he was being packed up and made ready for shipping at the close of SXSW Interactive. A reminder that AI needs a break just like the rest of us.