Introduction
Last year, I watched a video of Alex Conley, a man with a cervical spinal cord injury, controlling a robotic arm mounted to his wheelchair. Not with a joystick or a keyboard. With his thoughts. A Neuralink brain implant translated what he imagined into motion: putting on a hat, opening a door, microwaving a pretzel and eating it on his own. Tasks he had not been able to do since his accident.
I keep coming back to that video because it captures something most conversations about AI completely miss. We have spent the last few years arguing about whether AI will write our emails or steal our jobs. Meanwhile, AI quietly learned to move things.
The Overlooked Piece of AI Innovation
When most people say “artificial intelligence,” they still mean software. The primary use cases are still chat bots, image generators, and coding agents. And yes, those tools are changing how we work. But they are also a distraction from something much larger.
With the recent release of Claude Cowork (not to mention the viral OpenClaw), the average corporate employee has gotten a small taste of what AI can do when its environment is expanded. When computer-using agents are given broader access to files, they can autonomously complete end-to-end work with minimal human intervention. But what happens when we give AI agents an even bigger sandbox, expanding their reach outside the confines of the digital world and into the physical world?
New products are being developed to bring AI into the physical world. Bionic hands that use AI to read nerve signals and grip objects the way a natural hand would. AR glasses that use AI to transcribe speech in real time, projecting captions into a deaf person’s field of vision during a live conversation. Industrial exoskeletons trained on billions of motion data points that learn each worker’s movement patterns and adjust support in real time. A surgical robot trained on videos of operations that performed a phase of a gallbladder removal without human help, responding to voice commands like a novice surgeon working with a mentor.
And the pace is accelerating. A recent study presented at the Augmented Humans International Conference found that the human brain can adapt to control robotic extensions, revealing what the researchers call “the brain’s adaptability to new augmentations.”1 Amazon has deployed over one million AI-powered robots across more than 300 warehouses worldwide, coordinated by a generative AI model that optimizes their movement in real time.
The world of AI has already begun expanding beyond the confines of a computer screen and into the physical world.
The Question That Changes Everything
Here is a question that made me think: “If an AI-powered exoskeleton moves before you intend it to, is it assisting or taking control?”1
Consider that for a second. The technology is getting good enough to anticipate what you need before you consciously ask for it. A robotic arm that steadies a surgeon’s hand mid-cut sounds like a miracle on paper.
But when the machine acts before the human decides, who is in charge?
That is not a philosophy seminar question. It is a question your legal team, your operations team, and your board will need to answer the moment you deploy any physical AI system. Who is liable when an AI-controlled forklift makes a decision a human operator wouldn’t have? What happens when an exoskeleton overrides its wearer’s movement to prevent an injury that may or may not have been about to happen?
The leaders figuring this out right now, before the technology forces the issue, will be the ones setting the terms for everyone else.
Why Human Creativity Becomes More Important, Not Less
I have spent my career studying human creativity, building companies, and investing in startups. And I will tell you the same thing I tell the audiences I speak to: the more AI can do, the more your distinctly human skills matter.
When a machine can lift the heavy boxes or guide the scalpel, the competitive edge moves to the person who decides what to build, why it matters, and how to make people care about it. Creativity, innovative thinking, and the willingness to challenge the way things have always been done will be the biggest differentiators when routine physical and cognitive work gets automated.
Three Things Every Leader Should Do Right Now
1. Stop treating AI as only a screen problem. Most companies are still asking “how do we use AI in our software?” Start asking an additional question: “where are our people doing physical work that AI could augment?” Walk your factory floor. Tour your warehouse. Shadow your field technicians. The biggest AI opportunities in your organization might be outside of your IT department.
2. Design with your people, not for them. The research on augmentation technologies is consistent on this point: the systems that work are the ones built collaboratively with the humans who use them.1 Handing someone an AI-powered tool and expecting them to trust it is like handing someone a parachute and pushing them out of a plane. Involve them in the design. Let them shape how it works. Trust comes from participation, not from a memo.
3. Build the guardrails before you need them. Questions about control, liability, and access are coming. You can answer them on your own terms now, or you can answer them in a courtroom later. Do not be the company that figures out its ethical framework after the incident.
Frequently Asked Questions
Q: What does “AI shaping the physical world” actually mean?
It means AI is leaving the screen. Robotic exoskeletons, neural interfaces, haptic wearables, AI-guided surgical instruments: these are all technologies where AI directly interacts with the physical environment. Researchers describe it as AI, robotics, and extended reality “redefining human potential, enhancing cognition, perception, and physical abilities.”1
Q: Is this only relevant to manufacturing and healthcare?
Those industries are out front, but the applications go much further. Construction, logistics, defense, education, creative arts, remote work. If a job involves a human body doing something, physical AI will eventually change how it gets done.
Q: Will this replace workers?
The better word is augment. The most effective systems expand what a person can do rather than eliminate the person entirely. Amazon’s warehouse robots handle the heavy lifting and long-distance transport so human workers can focus on tasks that require judgment and dexterity. AI-powered bionic hands are giving amputees grip control they never had with traditional prosthetics. These technologies have the potential to create capability, not unemployment.1
Q: What are the biggest risks?
Loss of human control is number one. After that: security vulnerabilities in connected physical systems, unequal access (only wealthy organizations benefiting), and accountability gaps when an AI-powered device makes a bad call. All of these are solvable, but only if you take them seriously before something goes wrong.
Q: Where should a leader start?
Walk the floor. Identify the physical work your people do every day. Ask yourself which tasks could be safer, faster, or more accessible with AI-powered augmentation. Then talk to the people doing that work before you build or buy anything. The best technology decisions start with the humans, not the hardware.
Citations:1 Li, J., Withana, A., Diening, A., Kunze, K., & Inami, M. (2025). Beyond Human: Cognitive and Physical Augmentation through AI, Robotics, and XR. Workshop at the Augmented Humans (AHs) International Conference 2025. arXiv:2503.09987.