Pod Insights Short Wave Is AI Ready For Robots? About this episode It seems like artificial intelligence is everywhere in our virtual lives. It's in our search results and our phones. But what happens when AI moves out of the chat and into the real world? NPR science editor and correspondent Geoff Brumfiel https://www.npr.org/people/279612138/geoff-brumfiel took a trip to the Intelligence through Robotic Interaction at Scale Lab at Stanford University to see how scientists are using AI to power robots and the large hurdles that exist for them to perform even simple tasks.
Read Geoff's full story https://www.npr.org/2025/03/17/nx-s1-5323897/researchers-are-rushing-to-build-ai-powered-robots-but-will-they-work.
Interested in more AI stories? Email us your ideas at [email protected] mailto:[email protected] .
See pcm.adswizz.com https://pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.
NPR Privacy Policy https://www.npr.org/about-npr/179878450/privacy-policy
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Speaker 00 This message comes from Doppel. Social engineering attacks come in disguise, but Doppel sees through them. The Agentic platform shuts down threats, turns employees into first line defenders, and more. Learn more at D-O-P-P-E-L dot com.
0:16 Speaker 01 You're listening to Short Wave from NPR.
0:21 Regina Barber? Hey, Short Wavers. Regina Barber here. It seems like artificial intelligence is everywhere in our virtual lives. It's in our search results, our phones. It's trying to read my emails. But NPR science correspondent Jeff Brunfield has noticed that AI isn't just showing up online anymore, it's starting to creep into reality.
0:39 Speaker 03 Yep. I don't know if you tuned in for Tesla's big marketing event last year, Regina.
0:44 Regina Barber? No.
0:44 Speaker 03 But AI was there.
0:47 Speaker 04 Speaking of robots.
0:50 Speaker 03 Tesla is obviously a car company.
0:53 Regina Barber? Yep.
0:53 Speaker 03 But Elon Musk, Tesla's CEO, made a big part of the event about a humanoid robot powered by AI and called Optimus.
1:02 Speaker 04 The software, the AI inference computer, it all actually applies to a humanoid robot.
1:09 Speaker 03 And Google just unveiled another humanoid robot that operates using AI.
1:14 Speaker 05 We're bringing Gemini 2.0's intelligence to general purpose robotic agents in the physical world.
1:21 Regina Barber? OK, Jeff, but even before AI came along, people and companies have been making like big claims about robots.
1:27 Speaker 03 They have, they have. And the robots, as I'm sure you know, Gina, have always disappointed compared to the vision.
1:32 Regina Barber? Yeah, that's true.
1:33 Speaker 03 And that's why I set out to understand the truth about AI and robotics.
1:37 Regina Barber? The truth.
1:38 Speaker 03 And I think I kind of found it in a bowl of trail mix.
1:45 Regina Barber? Today on the show, what happens when artificial intelligence moves out of the chat and into the real world?
1:51 Speaker 03 We're looking at how AI could maybe revolutionize robotics.
1:55 Regina Barber? You're listening to Short Wave, the science podcast from NPR.
2:06 Speaker 06 Stay tuned for the following announcements and messages from our sponsors.
2:10 Speaker 00 This message comes from Avalara Aviator. Aviator is the agent hub for tax and compliance. If you've listened to enough podcasts, you know a confident answer isn't necessarily a correct one. Sources matter. The same goes for AI, tax, and compliance. Aviator is grounded in 20-plus years of Avalara tax intelligence and delivers source-backed, audit-defensible outcomes with human oversight built in. Because the AI said so, won't pass an audit. Visit avalara.com/aviator.
2:40 Speaker 07 Oh, Kevin, it feels good to be back. It sure does, Casey. And what a time. The world is talking about AI and what's going on in Silicon Valley. And we are here with a new show to tell you all about it. That's right, Kevin. On our new show, Machine Gods, twice a week, we'll be bringing you the biggest news in tech and bring you unmatched access to the real titans of this industry. You can find Machine Gods from NPR on the NPR app or wherever you get your podcasts.
3:05 Speaker 01 This midterm election season, the NPR app is your single stop for clear, independent coverage. Stay on top of national and local storylines with news, stories, podcasts, and political analysis on demand. Download the NPR app and check out the Elections 2026 tab today.
3:29 Regina Barber? OK, so Jeff, you were interested in finding out more about how AI works in robots. Where did you start?
3:34 Speaker 03 Well, I didn't go to Tesla or Google, but I did drive right by them on my way to Stanford University.
3:39 Regina Barber? OK.
3:40 Speaker 03 And specifically, the IRIS Laboratory, which stands for Intelligence Through Robotic Interaction at Scale. I got a tour from a graduate student named Mu Jin Kim. Mu Jin works on a new kind of robot powered by AI, similar to the AI used in chatbots.
3:56 Speaker 05 It's one step in the direction of ChatGPT for robotics, but... Um, still a lot of work to do.
4:02 Speaker 03 Okay, all right. Well, you want to show me how it uh show me what I can do?
4:05 Speaker 05 For sure. So.
4:06 Regina Barber? So Jeff, what did the robot look like?
4:09 Speaker 03 Well, this wasn't some humanoid robot that the big tech companies are rolling out. It's just a pair of mechanical arms with pinchers.
4:16 Regina Barber? Okay.
4:17 Speaker 03 But what made it interesting was that it's powered by an AI model called OpenVLA. So first, we should probably just say quickly, you know, a regular robot must be very, very carefully programmed. An engineer has to write it detailed instructions for every task you want it to perform.
4:33 Regina Barber? Yeah, and AI is supposed to change that.
4:35 Speaker 03 Exactly. And that's what's going on here. This robot is powered by a teachable AI neural network. The neural network operates kind of how scientists think the human brain might work. Basically, there are these mathematical nodes in the network that have billions of connections to each other in a way similar to how neurons in the brain are connected together. And so when you go to program this sort of thing, it's simply about reinforcing the connections that matter between the nodes and weakening the other ones that don't. So in practice, this means Mu Jin can just teach OpenVLA a task by showing it.
5:10 Speaker 05 So basically, like whatever task you want to do, you just keep doing it over and over, maybe like like 50 times or 100 times.
5:17 Speaker 03 The robot's AI neural network becomes tuned to that task, and then it can do it by itself.
5:22 Regina Barber? Yeah, it makes me think of this like um smiling robot story we did. And that robot just watched like a lot of videos of people smiling, then it learned how to do it.
5:30 Speaker 03 Yeah, it's exactly the same thing, except instead of just smiling, this robot's actually doing stuff.
5:36 Regina Barber? Right.
5:36 Speaker 03 So to show me, Mu Jin brought out a tray of different kinds of trail mix, and I typed in what I wanted it to do.
5:42 Speaker 05 Okay, so scoop some green ones with the nuts into the bowl.
5:46 Regina Barber? Oh.
5:47 Speaker 05 See what happens.
5:49 Regina Barber? Okay, so Jeff, personally, I've been waiting for something like AI in robotics because you can teach it to do something, you can ask it to do something to like make me an ice cream sundae or something without like any fancy programming or special knowledge.
6:00 Speaker 03 That's exactly it, you know? And this really is the dream of the researcher who runs this laboratory. Her name is Chelsea Finn.
6:07 Speaker 01 So in the long term, we want to develop software that would allow the robots to operate intelligently in any situation.
6:14 Speaker 03 And by intelligently, she means the robot could understand a simple command like scoop some green ones into a bowl, or make me a sundae, and then execute in the real world.
6:23 Regina Barber? Mm hmm.
6:24 Speaker 01 Even just to do very basic things, like being able to make a sandwich, or being able to clean a kitchen, or being able to restock grocery store shelves.
6:31 Speaker 03 These are simple tasks that could help humans do their jobs, or do tasks at home. Now, Chelsea also has co-founded a startup called Physical Intelligence. It recently demonstrated a mobile robot that could take laundry out of a dryer and fold it. Again, this robot was taught by humans, training its powerful AI program.
6:52 Regina Barber? Okay, so ice cream sundaes, is that too advanced? Is folding an easier start?
6:57 Speaker 03 I mean, I'd actually argue, Gina, that folding is harder.
7:01 Regina Barber? Okay.
7:01 Speaker 03 Let me show you a video.
7:03 Regina Barber? Okay, it's going to the dryer, it's pulling stuff out, putting it in a basket. It has the concentration I have when I'm going to do laundry. It almost looks annoyed with folding, like I do. Oh my god, it's doing really well, actually.
7:19 Speaker 03 Yes, it is, right? And this is a complicated task. It's got to pull these clothes out, it's got to figure out what they are, and...
7:26 Regina Barber? It doesn't even have a head, but I'm like giving it personality. It looks like it's like, oh, I just got to fold another one. Okay, so is it really as simple as like just teaching a robot like what to do? Because if it was, wouldn't these robots be everywhere?
7:42 Speaker 03 Yeah, I mean, right? It looks cool on the video. The truth is that, you know, when you get out and these robots are trying to do these tasks over and over again, they get confused, they misunderstand, they make mistakes, and they just get stuck. So, you know, it might be able to fold laundry 90% of the time, or 75% of the time, but the rest of the time, it's going to make a big mess that then a human has to get in there and clean up.
8:06 Regina Barber? Got it, okay.
8:07 Speaker 03 I spoke to Ken Goldberg, professor at the University of California at Berkeley, and he is pretty emphatic that AI-powered robots weren't here yet. Robots are not going to suddenly become...
8:17 Speaker 04 The science fiction dream overnight.
8:20 Regina Barber? OK, so like, tell me why, because like AI chatbots have gotten like way better, super fast. So why are these robots getting stuck?
8:27 Speaker 03 OK, so it's true that AI has improved massively over the past couple of years. But that's because chatbots have a huge amount of data to learn from. They've taken basically the entire internet to train themselves how to write sentences and draw pictures. But Ken says.
8:43 Speaker 04 For robotics, there's nothing. We don't have anything to start with, right? There's no examples online of robot commands being generated in response to robot inputs.
8:55 Speaker 03 And if robots really need as much training data as their virtual chatbot friends, then having humans teach them one task at a time is going to take a really long time.
9:05 Speaker 04 You know, at this current rate, we're going to take 100,000 years to get that much data.
9:09 Regina Barber? What? OK, that's so long. Like, are there any alternatives? There must be.
9:14 Speaker 03 Yeah, well, scientists are exploring them right now. And one might be to let the AI brain of the robot learn in a simulation. A researcher who's trying this is a guy named Poolkit Agrawal. He's at the Massachusetts Institute of Technology.
9:30 Speaker 05 The power of simulation is that we can collect, you know, very large amounts of data. For example, in three hours, you know, worth of simulation, we can collect 100 days worth of data.
9:41 Speaker 03 So this is a really promising approach for some things, but it's much more of a challenge for others. So for example, let's talk about walking. When you're just dealing with the earth and your body, the physics of walking around, it's actually kind of simple.
9:55 Speaker 05 When you're doing locomotion, you know, you're mostly on earth. You know, there's no amount of force we can apply which will make the earth move.
10:02 Speaker 03 And so the simulation can do that reasonably well. But if you want your robot to say, try and pick up a mug off a desk or something, that's a lot more complicated.
10:11 Regina Barber? More forces.
10:11 Speaker 05 You know, if you apply the wrong forces.
10:13 Speaker 06 These objects and fly away very quickly.
10:16 Speaker 03 Basically, your robot will fling things across the room if it doesn't understand the weight and the size of what it's carrying. And there's more, you know, if your robot encounters anything that you haven't simulated 100% perfectly, then it won't know what to do. It'll just break.
10:31 Regina Barber? Okay, so it sounds like these like simulations have limits and real world training is going to take like a while. I can begin to see why AI robots aren't going to like be here tomorrow.
10:41 Speaker 03 Exactly. And some researchers think there are even deeper problems, actually, with trying to put AI into robotics. One of them is Matthew Johnson-Roberson at Carnegie Mellon University in Pittsburgh.
10:51 Speaker 04 In my mind, the question is not, do we have enough data? It is more, what is the framing of the problem?
10:59 Speaker 03 So getting back to AI chatbots for a minute. Matt says, for all their incredible skills, the task we're asking them to do is actually relatively simple. You know, you look at what a human user types and then try to predict the next words that user wants to see. Robots have so much more that they're going to have to do than just compose a sentence.
11:18 Regina Barber? Right.
11:19 Speaker 04 Next best word prediction works really well and is a very simple problem because you're just predicting the next word. And it is not clear right now, I can take 20 hours of GoPro footage and then produce anything sensible with respect to how a robot moves around in the world.
11:33 Speaker 03 So in other words, the sci fi tasks that we want our robots to do are so complicated compared to sentence writing. No amount of data may be enough unless researchers can find the right way to teach the robots.
11:46 Regina Barber? Or have the robots teach the robots.
11:48 Speaker 03 Yes, that's also an option. They can teach themselves.
11:51 Regina Barber? Okay, so Jeff, you've taken me from like optimist to pessimist. It's the, you know, the road I take every day. I'm starting to think that AI is like never going to work that well in robots or like it's going to be a really long time.
12:05 Speaker 03 You know, I'm sorry if I've like turned you into a pessimist here, Gina.
12:11 Regina Barber? It happens.
12:12 Speaker 03 And I'm going to have to sort of whipshaw you back because. AI is already finding its way into robotics in ways that are really interesting. So for example, Ken Goldberg has co-founded a package sorting company. And just this year, they started using AI image recognition to pick the best points for their robots to grab the packages.
12:32 Regina Barber? Ooh, OK.
12:33 Speaker 03 Yeah, and it's working really well, he told me. And I think we're going to see a lot of that, AI being used for parts of the robotic problem, you know, walking or vision or whatever. It's going to make big progress. It just may not arrive everywhere all at once. And to really end on a high note here, let's get back to that Stanford lab. Remember, I asked it to grab some trail mix, right?
12:56 Regina Barber? Yeah, yeah.
12:56 Speaker 03 So the robot correctly identified the right bin to Mu Jin Kim's relief.
13:02 Speaker 05 Usually that that spot right there where it identifies the object and goes to it, that's that's the part where we hold our breath.
13:08 Speaker 03 And then very, very slowly and kind of hesitantly, it reached out with its claw and picked up the scoop.
13:18 Regina Barber? It's doing it.
13:18 Speaker 05 Mu Jin, did I just program a robot?
13:22 Speaker 05 You did. Looks like it's working.
13:24 Speaker 03 And to my mind, it's incredible. Like, remember, nobody really programmed the robot exactly. This is all neural network learning how to move the claws and respond to the commands on its own. And to me, it's pretty wild that that works at all, and I think it's going to lead to some very cool developments.
13:41 Regina Barber? I'm excited to hear more, Jeff. Thank you so much for bringing this reporting to us.
13:45 Speaker 03 Thank you very much.
13:47 Regina Barber? We'll link Jeff's full story, which has robot videos in our episode notes.
13:53 Regina Barber? This episode was produced by Burley McCoy, edited by our showrunner Rebecca Ramirez, and fact-checked by Tyler Jones. Jimmy Keely was the audio engineer. Beth Donovan is our senior director, and Colin Campbell is our senior vice president of podcasting strategy.
14:06 Speaker 03 I'm Jeff Brumfield.
14:07 Regina Barber? I'm Regina Barber. Thank you for listening to Short Wave from NPR.
14:21 Speaker 00 Stay tuned for the following announcements and messages from our sponsors.
Previous Rare Narwhal Footage Shows New Tusk Activities Next The Iguanas That Rafted To Fiji Page updated November 3, 2026 · transcript and summary are machine-generated