We prefer robots with expressions, even if they’re less efficient

Experimental, Identity, Personality, Research, Robotics, Self

Monday, August 22nd, 2016

What’s the one thing we all love more than anything else? Ourself.

We make our ways through the rat race, zigging and zagging around corners, trying to hold on to our fleeting lives so our selves can see another day…and when we picture living that next day, we most likely see ourself as our physical body moving through time and space.

You’ll picture yourself driving to work, taking a hike, or playing with your dogs, all as your body. Thinking this way only strengthens the natural affinity you have toward yourself and the human aesthetic.

Beyond that, we’re so ego-driven we see faces in everyday objects (pareidolia), and we anthropomorphize animals to be more like us. The facts are, we love people (at least most of us), and its this love that’s allowed us to grow into what we are as a society.

We love working as groups, and if someone looks more similar to us, guess what we do? We like them more.

But exactly how far does this idea of self-embellishment extend?

Adding Personality to Robots

We’re making our way into a new, more futuristic era, and terms like robotic ethics are becoming more common. In fact, integration between robots and humans, especially in the workplace, has already started.

I went on a brewery tour at a local craft beer company about a month or so ago (insert mock hipster comments now). As our time was ending, the tour guide asked if we wanted to go to the back and see something they normally don’t let tours see. Of course, the group collectively agreed.

As we made our way through several corridors packed to the ceilings with beer, we came to a large, looming black cage in the back that seemed almost out of place.

Inside the cage was a behemoth of a bright yellow machine the staff referred to by name. Its arm extended out like a crane, giving it the ability to lift kegs with ease. As we all stared in fascination, the guide went on to explain that the average life for a keg-moving employee was about 3 years. Every one of their backs were eventually giving out because it’s just not natural to move kegs around all day.

“With the newest addition to the team, the crew can have him do the lifting at little expense to them. The robot’s just great.” he said as he wrapped it up.

“We all just love him around here, and he’s become more like family, especially to the guys out here working with him.”

The Research

A recent press release from the University of Bristol explored human-robot interaction and how we come to put our trust in our future co-droid-workers.

More specifically, they looked at how integrating humanlike facial expressions would affect people’s perceptions of their robot counterparts. They wanted to see if these expressive features would mitigate people’s feelings, even if the robot was less efficient.

So they set out to make omelets. That’s right, omelets.

They set up an experiment with three different robots they dubbed BERT A, BERT B, and you guessed it, BERT C. The BERT‘s would each make an omelet for participants, but they had some variability.

BERT A was the most efficient, but completely uncommunicative.

BERT B was also non-communicative, but also made a mistake during the egg-making process.

BERT C also made a mistake during the omelet procurement, but was completely communicative and expressive once doing so. After fudging things up a bit, (I’m using gender loosely here) he’d look at participants, acknowledge the error, and apologize with a sad face.

Well, What Happened?

The team analyzed pre- and post-experiment questionnaires alongside some behavioral data recorded during the experiment when the robot would make the mistake.

As you may expect, no one liked BERT B. He’s the friend no one wants around, and everyone talks about behind their back. An error with no care or responsibility taken just doesn’t lead folks to like ya.

While there was a significant difference between BERT B and his A and C counterparts, there was no reliable difference in how people felt toward BERT A and BERT C.

However, people rated BERT C more favorably than both A and B. When BERT C interacted with his peers, some people were seen to actually mimic his expressions, enticing the thought of emotional contagion at work.

Although that thought is exciting, there’s no real idea if it’s empathy, sympathy, or the aforementioned at play here. What we do know is this is exciting news. It provides an unparalleled look at how social interaction drives us as a species and gives us a good look at why we may be seeing more humanlike robots in the future.

Wrapping it up

Anyone designing any robotic device can learn something from this study. The more human you make it, odds are, the consumers will like it more.

We love ourselves, and we can more easily empathize with others like us.

We just understand them better. If something takes credit for a mistake and apologizes, it can be forgiven. Not only that, it could actually start building a relationship with it’s human peer, increasing their preference of the automaton.

On the other hand, if your robot just sits there letting eggs crack like an ass, ain’t nobody got time for that.


  1. ‘Believing in BERT: Using expressive communication to enhance trust and counteract operational error in physical Human-Robot Interaction’. Hamacher, A., Bianchi-Berthouze, N., Pipe, A.G., & Eder, K. IEEE Conference Publications.

About Eric

I am the founder, designer, coder, writer, and jack-of-all-trades here at PsychMob. Living in beautiful Colorado, I spend most of my spare time hiking, meditating, and altering my consciousness. I'm relentlessly in love with psychology and pull no punches when I write about it. I hope you enjoy the site as much as I do! Thanks for coming.

Categories:  Experimental / Identity / Personality / Research / Robotics / Self

No Comments

Comments are closed.