Next up for AI? Dancing robots | Catie Cuan

[00:00:00] [SPEAKER_02]: TED Audio Collective

[00:00:10] [SPEAKER_01]: Soul singer and songwriter Erykah Badu travels around with her robot P-Funk.

[00:00:15] [SPEAKER_01]: It trails behind her carrying snacks, veggies from her garden,

[00:00:19] [SPEAKER_01]: clothes, pens for the laundry, whatever she needs,

[00:00:22] [SPEAKER_01]: and serving as her companion for errands and interesting social media fodder.

[00:00:27] [SPEAKER_01]: We are living in an age where robots are embedded in our daily lives.

[00:00:31] [SPEAKER_01]: This is taking shape in our most intimate of spaces.

[00:00:34] [SPEAKER_01]: Autonomous servers make your pizza or serve your coffee.

[00:00:37] [SPEAKER_01]: They become 24-hour law enforcement in public spaces.

[00:00:41] [SPEAKER_01]: They even help you cross the street.

[00:00:43] [SPEAKER_01]: We're only becoming more and more dependent on this technology.

[00:00:47] [SPEAKER_01]: So as we look to the future,

[00:00:49] [SPEAKER_01]: how can we build more empathetic robots that seamlessly integrate into our lives?

[00:00:55] [SPEAKER_01]: This is TED Tech, a podcast from the TED Audio Collective.

[00:01:00] [SPEAKER_01]: I'm your host, Sherelle Dorsey.

[00:01:01] [SPEAKER_01]: Today on the TED stage, we hear from choreographer and roboticist Catie Cuan.

[00:01:09] [SPEAKER_01]: She teaches robots how to dance.

[00:01:12] [SPEAKER_01]: Her ultimate goal is to build humanity into our robotic technology

[00:01:16] [SPEAKER_01]: by training it to move just like people do.

[00:01:21] [SPEAKER_01]: But before we dive into the talk, a quick break to hear from our sponsors.

[00:01:36] [SPEAKER_03]: Imagine this. In 2030, the CFO of a Fortune 100 company is a bot.

[00:01:41] [SPEAKER_03]: I'm Paul Meikleman, and on Imagine This, we'll be exploring possible futures

[00:01:45] [SPEAKER_03]: and the implications they hold for organizations.

[00:01:49] [SPEAKER_03]: Joining me will be BCG's top experts, as well as my co-host, Gene,

[00:01:53] [SPEAKER_02]: BCG's conversational gen AI agent.

[00:01:55] [SPEAKER_02]: Blending human creativity with AI innovation,

[00:01:58] [SPEAKER_02]: this podcast promises an unmatched listening journey.

[00:02:02] [SPEAKER_02]: Join us on Imagine This from BCG.

[00:02:08] [SPEAKER_05]: Sometimes things in the world of technology are complicated

[00:02:11] [SPEAKER_05]: and need careful explaining.

[00:02:14] [SPEAKER_05]: Sometimes they just need a little hard truth.

[00:02:17] [SPEAKER_00]: I don't think anyone is going to buy a banana with crypto

[00:02:20] [SPEAKER_00]: at any point in the foreseeable future.

[00:02:23] [SPEAKER_05]: I'm Lizzie O'Leary, the host of Slate's What Next? TBD,

[00:02:26] [SPEAKER_05]: your clear-eyed guide to technology, power, and the future,

[00:02:29] [SPEAKER_05]: Friday and Sunday, wherever you get your podcasts.

[00:02:32] [SPEAKER_04]: I have an unusual job.

[00:02:39] [SPEAKER_04]: I teach robots how to dance.

[00:02:41] [SPEAKER_04]: In 2017, I transitioned from dancing on stages with humans

[00:02:47] [SPEAKER_04]: to dancing with robots.

[00:02:51] [SPEAKER_04]: Big robots, small robots, single arms, and roving robots.

[00:02:56] [SPEAKER_04]: Robots that fly, vacuuming robots,

[00:02:59] [SPEAKER_04]: a lot of robots that look like humans,

[00:03:01] [SPEAKER_04]: and a whole bunch that don't.

[00:03:03] [SPEAKER_04]: I'm pretty sure I've choreographed and danced

[00:03:06] [SPEAKER_04]: with more different types of robots than anyone on Earth.

[00:03:09] [SPEAKER_04]: Why do I do this?

[00:03:11] [SPEAKER_04]: Because robots move,

[00:03:13] [SPEAKER_04]: and choreographers understand how movement affects us.

[00:03:17] [SPEAKER_04]: If a robot slides politely out of a doorway to let you pass,

[00:03:21] [SPEAKER_04]: it might make you feel seen and acknowledged.

[00:03:24] [SPEAKER_04]: If a robot marches quickly towards you

[00:03:26] [SPEAKER_04]: and avoids you at the last second,

[00:03:28] [SPEAKER_04]: it might cause revulsion and fear.

[00:03:31] [SPEAKER_04]: Robots are beginning to show up in our everyday environments,

[00:03:34] [SPEAKER_04]: from sidewalks to offices, backyards to hospitals,

[00:03:37] [SPEAKER_04]: and they will be threatening and confusing to us

[00:03:40] [SPEAKER_04]: if we do not carefully examine how they move.

[00:03:45] [SPEAKER_04]: I'm also a roboticist and an engineer.

[00:03:48] [SPEAKER_04]: I study this connection between robot motion and emotion.

[00:03:52] [SPEAKER_04]: I combine my expertise in robotics and dance

[00:03:55] [SPEAKER_04]: and the emerging field of choreo-robotics,

[00:03:59] [SPEAKER_04]: the intersection of dance and robotics.

[00:04:02] [SPEAKER_04]: Right now, we're giving AI a robot body

[00:04:06] [SPEAKER_04]: with physical intelligence.

[00:04:08] [SPEAKER_04]: AI technologies like GPT and Gemini

[00:04:11] [SPEAKER_04]: are becoming the robots' brains,

[00:04:13] [SPEAKER_04]: which means that robots are getting smarter,

[00:04:17] [SPEAKER_04]: much smarter.

[00:04:18] [SPEAKER_04]: Think of what AI has already done to text and video generation.

[00:04:22] [SPEAKER_04]: We are just scratching the surface

[00:04:24] [SPEAKER_04]: on seeing what AI can do in the physical world.

[00:04:28] [SPEAKER_04]: We are going to be interacting with these new intelligent robots

[00:04:32] [SPEAKER_04]: more and more,

[00:04:33] [SPEAKER_04]: interacting with them through movement.

[00:04:38] [SPEAKER_04]: Roboticists already spend a lot of time thinking about movement,

[00:04:41] [SPEAKER_04]: but they think about it in terms of efficiency,

[00:04:44] [SPEAKER_04]: optimizing for speed or success.

[00:04:47] [SPEAKER_04]: That's not enough.

[00:04:49] [SPEAKER_04]: We need to teach robots how to dance.

[00:04:53] [SPEAKER_04]: If a robot can dance,

[00:04:54] [SPEAKER_04]: make every movement with intention, agility, balance,

[00:04:59] [SPEAKER_04]: emotional expression and grace,

[00:05:01] [SPEAKER_04]: I believe they will also become better caregivers,

[00:05:04] [SPEAKER_04]: teachers and companions.

[00:05:06] [SPEAKER_04]: It will help us build better general-purpose robots

[00:05:09] [SPEAKER_04]: that we actually want to live amongst us.

[00:05:13] [SPEAKER_04]: If we teach robots to dance,

[00:05:15] [SPEAKER_04]: we expand the future of intelligent machines.

[00:05:21] [SPEAKER_04]: Choreographers and engineers have made dances with robots

[00:05:24] [SPEAKER_04]: since the 1990s,

[00:05:25] [SPEAKER_04]: and these highly scripted performances

[00:05:27] [SPEAKER_04]: have captured the public's imagination.

[00:05:31] [SPEAKER_04]: They've made us feel wonder,

[00:05:33] [SPEAKER_04]: made us laugh,

[00:05:34] [SPEAKER_04]: made us want to dance along,

[00:05:36] [SPEAKER_04]: and sometimes made us never want to meet a robot in real life.

[00:05:41] [SPEAKER_04]: Dancing with robots is no longer just about art or entertainment.

[00:05:46] [SPEAKER_04]: It is about transferring the lessons that,

[00:05:49] [SPEAKER_04]: one, all movement changes emotions,

[00:05:53] [SPEAKER_04]: and two, contextual movement is critically important

[00:05:57] [SPEAKER_04]: in applying those lessons to robots in our everyday lives.

[00:06:02] [SPEAKER_04]: Robotics engineers are beginning to consider emotion and context

[00:06:07] [SPEAKER_04]: when designing intelligent machines that interact with us.

[00:06:11] [SPEAKER_04]: The choices we make now

[00:06:13] [SPEAKER_04]: are going to dictate how we live our lives with robots.

[00:06:19] [SPEAKER_04]: If we get this right,

[00:06:20] [SPEAKER_04]: robots will be more welcomed, safe, and delightful.

[00:06:25] [SPEAKER_04]: Humans will be more empowered and comfortable.

[00:06:29] [SPEAKER_04]: Before AI, programmers needed hours

[00:06:32] [SPEAKER_04]: to script a simple dance sequence for a robot to perform,

[00:06:36] [SPEAKER_04]: just like they needed hours to script the robot

[00:06:39] [SPEAKER_04]: to open a single door.

[00:06:41] [SPEAKER_04]: With AI, you can teach the robot

[00:06:43] [SPEAKER_04]: to open just a few specific doors,

[00:06:46] [SPEAKER_04]: and it will learn to open all of them,

[00:06:48] [SPEAKER_04]: even ones it hasn't seen before.

[00:06:51] [SPEAKER_04]: It's also true for dance.

[00:06:53] [SPEAKER_04]: You can teach the robot to dance with a specific person,

[00:06:57] [SPEAKER_04]: and it will learn how to dance and move with many others

[00:07:01] [SPEAKER_04]: in many different environments and circumstances.

[00:07:04] [SPEAKER_04]: This is what I did at Everyday Robots,

[00:07:07] [SPEAKER_04]: then a robot AI moonshot at Google.

[00:07:10] [SPEAKER_04]: But rather than teach one robot,

[00:07:13] [SPEAKER_04]: I used AI to teach 15 robots how to move together as a flock.

[00:07:19] [SPEAKER_04]: We imagined a world where you could walk down a hallway

[00:07:22] [SPEAKER_04]: filled with robots,

[00:07:24] [SPEAKER_04]: and they would part to make space for you,

[00:07:27] [SPEAKER_04]: like a flock of doves or a crowd of people on a city street,

[00:07:31] [SPEAKER_04]: where a robot could navigate seamlessly and even beautifully

[00:07:35] [SPEAKER_04]: through a busy, chaotic Times Square.

[00:07:38] [SPEAKER_04]: The world we live in is complicated

[00:07:41] [SPEAKER_04]: for robots to understand,

[00:07:43] [SPEAKER_04]: so it was impossible to script or choreograph

[00:07:45] [SPEAKER_04]: interactions like these step by step.

[00:07:48] [SPEAKER_04]: We needed to use the magic of AI

[00:07:50] [SPEAKER_04]: to teach the robots how to improvise and react.

[00:07:54] [SPEAKER_04]: This was a massive challenge that had never been done before,

[00:07:58] [SPEAKER_04]: and the result is an interactive performance

[00:08:00] [SPEAKER_04]: and installation work called Project Starling.

[00:08:04] [SPEAKER_04]: This took us two years of hardcore engineering to complete.

[00:08:08] [SPEAKER_04]: We trained an AI agent on the preferences of a choreographer, me.

[00:08:13] [SPEAKER_04]: I chose the most engaging patterns and sequences.

[00:08:17] [SPEAKER_04]: We also taught the robots how to recognize

[00:08:20] [SPEAKER_04]: and respond to human gestures.

[00:08:24] [SPEAKER_04]: While we were at it,

[00:08:25] [SPEAKER_04]: we turned the robots into musical instruments.

[00:08:29] [SPEAKER_04]: Each joint on the robot was mapped to a different sound.

[00:08:33] [SPEAKER_04]: When the robot's torso moved, it triggered a bass sound.

[00:08:38] [SPEAKER_04]: Its hand opening and closing triggered a bell sound.

[00:08:43] [SPEAKER_04]: The robots' actions created its own unique symphony.

[00:08:48] [SPEAKER_04]: I wanted it to feel like stepping out of your normal reality

[00:08:52] [SPEAKER_04]: and into a waking dream,

[00:08:55] [SPEAKER_04]: one that was so far from where we stand today,

[00:08:58] [SPEAKER_04]: yet all the same felt like a world

[00:09:01] [SPEAKER_04]: where your kids could grow up and feel calm,

[00:09:05] [SPEAKER_04]: where you could be surrounded by machines,

[00:09:08] [SPEAKER_04]: and it felt natural and safe.

[00:09:11] [SPEAKER_04]: People were drawn to just wander amongst these robots.

[00:09:15] [SPEAKER_04]: You know, some were a bit tentative at first,

[00:09:17] [SPEAKER_04]: but then they began to engage and explore

[00:09:20] [SPEAKER_04]: with curiosity and playfulness.

[00:09:23] [SPEAKER_04]: People kept coming back.

[00:09:26] [SPEAKER_04]: What we learned about gesture, navigation,

[00:09:29] [SPEAKER_04]: human interaction and expression

[00:09:31] [SPEAKER_04]: can be applied to robots more generally.

[00:09:35] [SPEAKER_04]: Think of the following future scenario.

[00:09:37] [SPEAKER_04]: A robot in a senior living facility

[00:09:39] [SPEAKER_04]: waiting in a crowded cafe.

[00:09:43] [SPEAKER_04]: You sit at your table, you wave over to it,

[00:09:46] [SPEAKER_04]: make a sipping gesture,

[00:09:48] [SPEAKER_04]: and the robot understands that you would like a glass of water.

[00:09:52] [SPEAKER_04]: And when it drives over to you with its glass of water,

[00:09:55] [SPEAKER_04]: it nods at the other people that passes

[00:09:58] [SPEAKER_04]: and pauses graciously to let them go by.

[00:10:02] [SPEAKER_04]: And when it arrives at your table,

[00:10:03] [SPEAKER_04]: it places the glass gracefully down.

[00:10:07] [SPEAKER_04]: And when you thank it, it nods in acknowledgement.

[00:10:11] [SPEAKER_04]: Maybe it even performs a small celebratory dance in place.

[00:10:16] [SPEAKER_04]: We can choose to build and program robots

[00:10:20] [SPEAKER_04]: that we welcome.

[00:10:22] [SPEAKER_04]: Robots that make us feel understood.

[00:10:26] [SPEAKER_04]: Robots that we want to spend time with.

[00:10:29] [SPEAKER_04]: The promise of robots that will live and work amongst us

[00:10:32] [SPEAKER_04]: has existed for a very long time.

[00:10:35] [SPEAKER_04]: From Carol Kopeck's 1920 play,

[00:10:39] [SPEAKER_04]: Rossum's Universal Robots,

[00:10:41] [SPEAKER_04]: to The Jetsons' Rosie the Robot,

[00:10:43] [SPEAKER_04]: to C3PO of Star Wars,

[00:10:45] [SPEAKER_04]: and even Terminator.

[00:10:47] [SPEAKER_04]: These stories have been deeply impactful

[00:10:50] [SPEAKER_04]: in what we imagine and build.

[00:10:54] [SPEAKER_04]: Yes, we can have robots that do our chores.

[00:10:57] [SPEAKER_04]: But we can also have robots be our dance partners.

[00:11:03] [SPEAKER_04]: Humans do a range of things.

[00:11:05] [SPEAKER_04]: Ceramics to diplomacy,

[00:11:07] [SPEAKER_04]: singing in choirs to vertical farming.

[00:11:10] [SPEAKER_04]: Imagine a world where robots can do a myriad of things

[00:11:14] [SPEAKER_04]: that are functional or creative,

[00:11:18] [SPEAKER_04]: and everything in between.

[00:11:21] [SPEAKER_04]: General purpose robots are coming,

[00:11:24] [SPEAKER_04]: and they will help us to address big demographic challenges

[00:11:27] [SPEAKER_04]: like aging populations and massive labor shortages.

[00:11:32] [SPEAKER_04]: And how they move will unlock

[00:11:35] [SPEAKER_04]: whether we let them into our homes

[00:11:38] [SPEAKER_04]: and welcome them in our workplaces.

[00:11:40] [SPEAKER_04]: Because we want to feel like we're still marvelously human

[00:11:45] [SPEAKER_04]: in this increasingly AI world.

[00:11:49] [SPEAKER_04]: And we can,

[00:11:50] [SPEAKER_04]: because we have the power to choreograph

[00:11:53] [SPEAKER_04]: the future we want.

[00:11:56] [SPEAKER_04]: Thank you.

[00:12:01] [SPEAKER_01]: That was Katie Kuhn at TED 2024.

[00:12:07] [SPEAKER_01]: TED Tech is part of the TED Audio Collective.

[00:12:10] [SPEAKER_01]: This episode was produced by Nina Lawrence,

[00:12:12] [SPEAKER_01]: edited by Alejandra Salazar,

[00:12:15] [SPEAKER_01]: and fact-checked by Julia Dickerson.

[00:12:18] [SPEAKER_01]: Special thanks to Maria Ladius,

[00:12:20] [SPEAKER_01]: Ferideh Granj,

[00:12:21] [SPEAKER_01]: Corey Hajim,

[00:12:22] [SPEAKER_01]: Daniela Valareso,

[00:12:24] [SPEAKER_01]: and Michelle Quint.

[00:12:25] [SPEAKER_01]: I'm Sherelle Dorsey.

[00:12:27] [SPEAKER_01]: Thanks for listening,

[00:12:28] [SPEAKER_01]: and talk to you again next week.