Picture this: You’re making cookies for a holiday get-together, and things have gotten hectic in the kitchen. You’ve opened the oven door, donned the oven mitts and grabbed a hot metal tray of warm snickerdoodles. You turn around to place them on the countertop and… whoops, you forgot to prepare something for the tray to rest on. As you weigh your options, you notice that some trivets have started to move out from their storage space on the counter. They’re rolling, on their own, right into place.
It seems like magic, like something out of Beauty and the Beast, but it’s one possible vision of your future kitchen, according to researchers at Carnegie Mellon University. With the help of cameras, a variety of AI models and some tiny little wheels, ordinary objects can find their way to the exact spot you wanted them to be, without you having to look for them.
It’s easy to picture a robot housekeeper, like Rosie from The Jetsons, but that isn’t the only way that robotics and artificial intelligence could theoretically make life easier for you at home or in the office. The same technology could be applied at a much smaller scale to the objects you already interact with regularly — your coffee mug, your stapler, your kitchen supplies and so on.
“Instead of bringing additional robots into our existing environments, what if the objects that are already there in our homes that we’re already familiar with can be both intelligent and robotic?” Violet Han, a Ph.D. student at CMU and lead author of a paper on the research, said in an interview.
Big, powerful humanoid robots give us a lot to worry about: They’re heavy and strong, capable of causing damage if they malfunction. They approach that uncanny valley of creepiness when something looks almost human. And it’s very hard to make one work reliably. Human dexterity is an extraordinary achievement of evolution, and we’ve built our world with the assumption that those who move in it can do things like grip a doorknob. That’s a tough skill to give to a robot. If those robots do become commonplace, they won’t be the only thing that’s automated.
“I have a hard time envisioning that you have these robot butlers, but then at the same time, everything else stays just as static as it was,” said Alexandra Ion, an assistant professor at CMU’s Human-Computer Interaction Institute, who leads the Interactive Structures Lab.
Adding AI and mobility to the objects we use solves many of those problems. It allows automation to feel more natural — you’re still using the same kind of stapler, even if it has little wheels and appears to have a mind of its own. But there are new problems, like privacy and security, that would have to be sorted out before your coffee mug starts to chase you every time you yawn.
Violet Han uses a stapler attached to a platform controlled by AI models.
Objects in motion
If you’re going to have smart trivets that know to roll over to you when you’re holding a hot tray of cookies, you need trivets that can move. For these experiments, the researchers built wheeled platforms — a round one for a mug, with a couple of wheels and motors and batteries, for example. Each one is operated by a Bluetooth-enabled microcontroller. A future consumer version of these objects could be custom-built to be a little more seamless, Han and Ion said, but this is just to prove the tech is capable.
The objects themselves are not equipped with artificial intelligence. While AI models can run on small pieces of hardware, like phones and watches, this is really one AI system controlling the whole kitchen.
It’s equipped with a camera that streams image frames to AI models that can process and identify what’s happening at a given moment. They determine what a person is doing and spot the objects that might be involved in that activity. Large language models with reasoning skills predict what is likely to happen next. A knowledge base encoded into the system ensures it knows some basic things about how people interact with objects.
“If a mug moves toward me, it’s more convenient for me if the handle is toward me,” for example, Han said.
While the idea of objects coming to save you in a crisis is dramatic (even if the crisis itself is fairly minor), the system could help in other ways. A key tray could shake your keys if you’re about to walk out of the house without them. If you’re looking for a stapler but it’s hidden behind something on your desk, the AI could move it to where you can see it. You could even use voice activation to ask your smart house to bring you the stapler.
When can you expect this in your house? The technology itself is “not that far off,” Ion said, but whether people would actually want the equipment that makes it possible is another matter. “If you are OK with overhead cameras, that’s much faster to deployment,” she said, “but personally, I wouldn’t be OK with that.”
One privacy solution is less technical than political. Better regulations and policies would give consumers the comfort that their privacy will be protected, Ion said. Having models that can run entirely on local hardware, computers not connected to the internet, would also help.
Watch this: How Humanoid Robots Can Earn Trust in 2026 | What the Future
What kind of robots do you want?
With humanoid robots, Ion said there’s a utopian view of android butlers and a “dystopian version where your robot butler might turn evil for some reason.” Even if the robot isn’t humanoid, there are still concerns. In this case, should you put AI-controlled wheels on a knife? The researchers had a moving knife, but designed it so that it would move with the blade always facing away from a person.
“I think it is an interesting tension and discussion to be had,” Ion said. “Do we just not want to have these types of objects at all being actuated?”
The goal is to ensure that when robots are doing things in our homes, they’re acting in a way that promotes safety and helps us actually do what we want to do.
“Robots are becoming increasingly capable of, for example, folding clothes, but… they should fold clothes in the way we want them to,” Han said. “Each piece of clothing may be different. It’s important for the robots not only to be capable but also to understand what the user wants and how they may best assist the users.”
One way that might look? Your coffee cup identifies that you’re ready for another sip and starts to find its way to you.
