What Does It Mean To Make a Robot Conscious?

THE FUTURE OF ROBOTS WITH HOD LIPSON

What does it mean to make a robot conscious? Can a machine have feelings? Will we make robots that take care of one another, upgrade themselves, and reproduce? I discussed these and other mind-bending questions with the brilliant founder of Columbia University’s Creative Machines Lab.

MEETING HOD LIPSON

I’ve wanted to meet Hod Lipson for a long time. A Professor of Mechanical Engineering and Data Science and founder and director of the Creative Machines Lab which pioneers both the ‘hard’ mechanics and the ‘soft’ AI underpinnings of robotics technology. He lives to push the boundaries of what robots can be. Hod has BIG ideas about the future, and he and his colleagues prove out those ideas with real-world, award winning prototypes.

We met at his lab, and as he walked me through the happy chaos of their daily lives: robot limbs, spare parts and experiments-in-progress, it struck me that not one experiment was trivial. Each represented a potential game-changer.

As you will hear, Hod took me in unexpected directions and introduced me to many new possibilities. I particularly enjoyed his knack for distilling complex ideas into simple one-liners. We covered plenty of ground and could have gone on for hours. All the laughter on the recording gives you some idea of all the fun we were having!

I hope you enjoy listening to our discussion as much as we did recording it.

CHECK OUT THE PODCAST TRANSCRIPT →

THE EXTRAORDINARY LIVES OF ROBOTS

If you’ve listened to the podcast, isn’t it interesting how everything Hod is working on, and I mean everything, links back to evolutionary principles? Robot bodies and minds, self-awareness and imagination, self-replication, soft robotics and soft actuators, robots that look after themselves and each other, robots that metabolize (eat!) other robots, a sustainable ‘ecology’ of robots!

In a sense, all his professional goals boil down to one: replicating the most powerful of nature’s mechanisms in a mechanical world. Evolutionary theory offers such a powerful, holistic lens through which to understand how the world works at every level. It also offers a powerful way to think not just about the future of robots but about the underpinnings of ALL technological and organisational innovation.

And how fascinating was the way robotic creativity and self-replication led Hod down a path to bioprinting and 3D food printing and (my favorite of all Hod’s ambitions) the 3D-printed robot that walks out of the printer?

ROBOT ECONOMICS 101

Also noteworthy is the close attention Hod pays to the economics. As he said, we cannot have a world where we make robots and throw them away. It simply isn’t sustainable. And as for robots manufacturing themselves: they manufacture everything else, so in what kind of future universe would robots NOT be manufactured by other robots?

ROBOT CONSCIOUSNESS

What does it mean for a robot to be conscious? The best moment for me was when Hod answered this in a phrase of just eight words. People still talk of consciousness as if it is a binary proposition – either an animal has it or it doesn’t – which is completely untenable.

Hod’s definition places it firmly on a continuum AND instantly helps us visualize the escalating complexity of consciousness – the rodent that sees itself eating a food scrap several seconds into the future versus the person visualising their financial situation and relationships many years into the future – as well as the exponentially escalating challenges in replicating the latter scenario. And exploring it through robot facial expressions and body self-awareness was just brilliant. The more I think about his definition the more powerful it gets!

AI ‘BLIND SPOTS’

Hod admitted to being blown away by the speed of advances in AI, which makes me think: when developments blindside someone like him, we’d better take notice! And he was genuinely troubled by the social dangers of conversational AI and robotics.

We disagreed on the timeframe for this one: based on my conversations with other AI scientists and technologists, I think fully-conversational AIs are coming much sooner than he does (our conversation was 4 months before the public launch of ChatGPT!). So what happens when the shy and socially-challenged start retreating into the embrace of fully-synthetic, humanoid robot companionships? It’s something to think about, isn’t it?

Remember that movie “Her” where Joaquin Phoenix falls in love with an AI virtual assistant voiced by Scarlett Johansson? Might be worth a re-watch.

RESPONSIBLE ROBOT SCIENCE

Hod is clearly an optimist who sees the benefits of his work outweighing than risks, but I really respect him for being so up front about his fears. After the recording, we talked about these some more.

He dwelled on the ‘double-edged’ possibilities of many of the technologies he is building. He then led me back to the roboticised head (pictured) and explained his latest ‘emotional engagement’ project.

The robot is designed to sense and predict when a human interlocutor is ABOUT to smile, and then to beat the human to it by a few hundred milliseconds. Robots that smile after we do, he pointed out, are perceived as imitators. Robots that smile first, are perceived as genuine, and there is a powerful difference in the emotional response and trust elicited in human interactions.

This is exciting territory. An abundance of positive applications await. But the potential for mis-use is also obvious, and Hod reiterated that he sees it as his scientific responsibility to communicate openly and frequently about his projects so we can better anticipate the challenges as well as the opportunities.

GROPING TOWARDS A 3D-PRINTED STEAK

When we got onto food, it was interesting to hear what Hod had to say about his very promising start with 3D-printed cakes and confectionary only to discover, as he diversified, that people are much more slavishly tied to the ‘familiar’ than he anticipated.

This applies to the entire preparation-cooking-eating experience – i.e. if a meat texture is perceived as odd in the uncooked state then it instantly loses appeal, no matter how good it might be cooked. 3-D printed food companies really have a job ahead of them.

I remain bullish specifically about combining 3D-printing with plant-based meat substitutes, and it was nice to learn that he agreed, and that, despite the challenges, his big ambition is adding the necessary structures and textures to boost plant-based meat products beyond meat patties! 

ROBOT GAME CHANGERS

Hod’s robots are already changing the world. They will continue to do so, I’m sure, for decades.

Thanks for hosting me, Hod. This visit was a joy. And don’t forget, I need to be there to eat that first delicious 3D-printed steak. I’m game.

And I MUST be there to witness that first 3D-printed robot walking out of the printer!