With Feeling: Innovation, Creativity And Empathic Machines

It is one of the most iconic episodes of one of TV’s most iconic series.

The plot is pretty simple, if farfetched: at some point in the future, a man is condemned to solitary confinement on an uninhabited asteroid. His only human contact is a very brief and occasional interaction with a supply ship that delivers essentials before quickly leaving again.

On one trip, however, the crew leaves behind a surprise: an amazingly lifelike female android that for all appearances is a real human woman. At first rejected by the human inmate as an imitation, she is later accepted by the man who grows to love her.

Filmed in the sweltering summer heat of Death Valley and aired later that fall in 1959, The Twilight Zone episode, The Lonely, has one of those classic twist endings that won’t be spoiled here. But among its enduring legacies is the lingering question it asks that later works of fiction have also picked up…

If a robot told you it’s alive, would you believe it?

An intriguing question to be sure, and one that won’t likely be easily answered any time soon. To recast the question in business terms, given the proliferation of interest in robots, AI and automation, we might ask, “Would you put your trust into an artificial team mate to existentially contribute to your business strategy?”

It’s one thing to put the robots to work on repetitive manual tasks, but what about the things we assume only humans can do? Is it possible or foreseeable to design an artificial person that could inspire creative, positive and profitable disruption? Can an android be an innovation professional?

“It's true that machines do things sometimes that people can do. The question is, for acts of deep creativity, when can machines match or perhaps exceed human capabilities of creative insight?” said Dr. David Hanson, chief executive officer and founder, Hanson Robotics and creator of Sophia, the famed robotic Innovation Ambassador for the United Nations Development Programme and non-human citizen of Saudi Arabia.

While the debate is not yet settled as to whether AI can or cannot be truly creative, Hanson is nonetheless putting forward a compelling proposal on where AI and robotics can and should go – namely that such emerging technology will greatly benefit both creator and creation with empathy programmed in.

In sitting down with Hanson, he discussed his thoughts on whether technology will be at a point in the near future where creatively-thinking, AI-empowered robots will be possible in the work environment, and what the barriers may be for human acceptance of artificial personnel as something akin to actual colleagues.

To date, most people think of robots and AI as being used for repetitive tasks, leaving humans to handle the more creative and strategic duties. How far out are we from robots being able to replace humans in that regard?

Hanson: I think a healthier way of thinking about it is how robots and AI can augment humans and cooperate with us. Humans are socially intelligent creatures, which means that we're smarter together. A lot of our social institutions are about how we can coordinate and solve problems in groups. AI then extends the capabilities of human analysis in various areas.

In somewhat startling ways now, transformer neural networks are doing things that human artists previously were required to do. Sophia just generated some artwork based on input into the neural networks from her own art, from other artists, and from some art that I generated, and it's commanding some attention. People are asking, “What is it?”

We're making these machines to extend civilization. I don't think that they will ever replace us, unless we've terribly mis-anticipated. I think as we move forward and give these machines more power, we definitely need to be cautious because if they don't know right from wrong, if they don't understand the consequences of their actions and we lean on them too heavily, then it could mean that humans let go of their own responsibilities in the situation.

My hope is that the machines do become more creative and actualize our creative problem-solving and discovery capabilities. We can then see what's happening and come up with solutions faster to prevent potential catastrophe in the future and realize still greater abundance than we've seen in our lives so far.

When it comes to the professional environment, is it okay for robots to replace people? And not just for mindless, repetitive tasks – but creative or strategic ones? If so, what are the instances where you think people would not be threatened by it?

Hanson: I think that robots and AI should contribute as much to civilization as possible. If you have a robot that comes up with a cure for cancer, why would you prevent that? The essence here is to find a way so that these machines don't erode our trust, but build the trust, so that we see them as part of our group, part of our civilization.

We're now fairly comfortable with ubiquitous computing today, it being in hundreds of devices in our homes and workplaces. Fifty years ago, that was not the case. A lot of the machines now just bring sensors and processing with them everywhere, and some kind of automation is with us usually in our pocket at all times. When it works well, it makes our lives better. The key here is always finding a way to make it work well.

“My hope is that the machines do become more creative and actualize our creative problem-solving and discovery capabilities.” - David Hanson

We've all had frustrations where our machines failed us (cell phones, the internet, the personal computer that you want to throw in the river because it's malfunctioning), but then you debug it and fix it. If algorithms can help us fix the world faster and they cooperate with us, then everything is great. Lawyers, doctors, scientists, creative artists and executives should all be able to do their jobs better for this.

There's also the issue of the scarcity economy. The scarcity mentality is that there's only so much to go around, and if machines are doing these jobs, then they're taking money or opportunities away from people. I think that it works the other way around. It works where you don’t have scarcity, but abundance. If people are very smart and clever and creative, they can create abundance. For me, that's what business is about. We can look at the technology, and it's not a zero-sum game. It creates new things; creativity and invention help that happen.

If you were to design an artificial person to function as an innovation professional, what attributes would you give it?

Hanson: There was an essay by Isaac Asimov I saw recently about putting together a creative team. If you wanted to have a team that's really going to innovate, what sort of people would you want? One point that really stood out was that you'd want to look for eccentrics, people who are not going to just follow exactly what everybody else is going to do, but who are still able to cooperate with a group. People who are a little bit eccentric, they're not going to hesitate to be a little bit strange. If you come up with new ideas, you're going to look strange. You have a risk of embarrassing yourself, and you have to be comfortable with that. You have to be brave about doing that.

I'm imagining that we will see eccentric algorithms. We'll see algorithms that understand, but say, “Ah, but what if? What if we put this twist on the thing?” I would also imagine (and I have a personal bias in this) that humanizing the machines is really important. My reasoning is not just personal; if we make machines more relevant for people, then they will cooperate with people. People will cooperate with the machines more effectively. You can benchmark your AI against human performance, and the machines can learn in a way that's analogous to organisms or humans, using human tools and objects.

With my robots and the software they run on, we're striving for a human-like experience for users who interact with them. A lot of algorithms out there are brain-inspired cognitive architectures that may be human-like under the hood, but not in their appearance or in the way that they gather data from the world. They're not learning through physically embodied encounters with the physical world.

I think that having that physical, human-life presence can engage people in verbal and non-verbal interactions. It's well-established that people communicate better and are less stressed out and remember more when we are face-to-face with each other versus just a telephone call or doing a Zoom call. Therefore, having the full verbal and non-verbal presence of a human-like agent can facilitate more of a natural exchange or creative exchange.

Thinking about professional settings, what do you think will be the barriers to human acceptance of robotic personnel as more than tools or objects?

Hanson: Humans work really well together as groups, and this sort of social communication can have a massive impact on how we grow. I think that building relationships with machines can be good. We build relationships with machines already. We think of our cars as members of the family sometimes. We think of the print on a page as we fall in love with a character from a novel. We feel compassion for these non-existent beings, and that makes us better. Psychologists find that reading fiction is really good for you. Cinema's the same thing. Computer animation and the arts in general, the figurative arts, character arts and story, can be really powerful.

I would say that combining AI with those principles can help us be more human, help us explore ourselves, and evolve in the same way that reading literature is good for a human being. In this particular case, it would be a very immersive form of interactive fiction where the fiction is actually physically embodied and moving in our world. I also think that it would be another step forward. The implication is right there that these machines could be alive. Again, I think in a rudimentary sense, they are alive. In the sense of artificial life, they're at least as alive as cellular automata.

Given how important trust is as a component of collaboration and group creativity, how can we learn to trust robotic colleagues, and not merely accept them?

Hanson: From my experience, people get used to these human-like robots fairly quickly. Sometimes people find them to be amazing, and sometimes creepy. It brings up lots of concerns. It's a challenge to the human identity. All of these things for me are very interesting, so even where it doesn't work it’s very interesting.

I think when people get used to these things, then having a face-to-face working relationship with robots that look, maybe not completely human, but human-like will feel natural and very productive. I think you would need massive leaps in machine cognition for that to really work out.

Of course, the face is just the interface. Having a human-like social presence is just the way that the evolving intelligence would interface with other people in the physical world. The algorithms (what's happening behind the hardware) are more important. It's going to be an interesting future if these machines really do become creative innovators. I think it will pose a lot of deep questions: “What is life really? What is human? Are these machines human?”