The Second Body: When Humans Inhabit Machines

The Second Body: When Humans Begin to Inhabit Machines

The next frontier of artificial intelligence may not be machines acquiring bodies, but humans acquiring additional ones.

In June 2025, Elon Musk made a remarkable suggestion during a Neuralink technology update. As the company's brain-computer interfaces become more advanced, he proposed, people might eventually control an entire Tesla Optimus humanoid robot using neural signals. Not simply a robotic hand or limb, but an entire body, complete with sensory feedback.

"So you could basically inhabit an Optimus robot," he said.

Inhabit. That is the interesting word.

We have grown accustomed to the idea that machines will become more like us. We give them intelligence, language, memory, reasoning and, increasingly, physical bodies. But Musk's proposition suggests movement in the opposite direction. What if humans become more capable of inhabiting machines?

What if the future is not simply about artificial intelligence becoming embodied, but about human intelligence becoming differently embodied?

And what happens to our understanding of identity when we can inhabit a second body, perhaps even a third, somewhere else in the world?

From controlling machines to inhabiting them

The immediate technological developments are significant, although they should not be mistaken for the full realisation of Musk's vision.

Neuralink has demonstrated that people with paralysis can use brain-computer interfaces to control computers and robotic devices. During the June 2025 presentation, the company showed an Optimus robotic hand moving through a replay of neural signals recorded from a participant.

But controlling a robotic limb is one thing. Experiencing that limb as part of your own body is something else.

There is already evidence that the distinction matters. In a 2021 study published in Science, researchers at the University of Pittsburgh demonstrated that artificial tactile feedback delivered through stimulation of the brain could substantially improve control of a robotic arm. A participant with paralysis was able to complete certain object-manipulation tasks in roughly half the time when touch sensations were added.

In other words, the machine was not simply receiving instructions from the person. The person was receiving sensory information back from the machine.

This is the beginning of something more interesting than remote control. It is the possibility of a two-way relationship between the human nervous system and an external mechanical body.

Of course, we are nowhere near demonstrating that an entire humanoid can be inhabited in the way Musk describes. Reliable whole-body movement, balance, proprioception, touch and the integration of sensory experiences present enormous challenges.

Nevertheless, the direction is worth examining.

Because the real question is not simply whether we can make a robot move by thinking.

It is whether we can come to experience that robot's body as our own.

The return of the body

For much of the digital era, we imagined the future of identity as a process of disembodiment.

We would become profiles, avatars, accounts, datasets and digital representations. Our social and professional lives would migrate into networks. We would increasingly exist in places where our physical bodies could not follow.

I have written extensively about this process through the idea of the Database Self: the person rendered machine-readable through accumulated information, behaviours and predictions.

But something else is now happening.

Artificial intelligence is moving out of screens and into the physical environment. Intelligence is becoming embodied in humanoid robots, embedded in infrastructure and increasingly autonomous in its actions.

In my recent keynote on the intelligent organisation, I explored these different forms of intelligence: unembodied, embodied, embedded and autonomous. I also examined neurotechnologies, including Neurable's EEG-based headphones, which demonstrate how technologies that sense patterns of brain activity are moving closer to the human body and nervous system.

These developments are usually discussed as advances in machine intelligence. What interests me increasingly is what they mean for human agency.

We have been asking what happens when machines acquire bodies.

Perhaps we should also ask what happens when bodies become available to human minds as new forms of interface.

A robot might no longer be something we merely instruct. It might become somewhere we experience ourselves as being.

That is a different relationship altogether.

When one self has several bodies

Imagine waking up in London but spending the first part of your morning inhabiting a humanoid robot in Tokyo.

You see through its cameras, move its hands, navigate its surroundings and receive some approximation of tactile sensation from its sensors.

Later, you disconnect and return your attention to your biological body.

In the afternoon, you inhabit a different robotic body to inspect machinery in a dangerous environment. Perhaps that body has capabilities your human body does not possess. It can withstand extreme temperatures, manipulate heavy objects or reach places inaccessible to human beings.

The scenario is speculative, but the underlying questions are already being investigated by researchers studying bodily self-consciousness.

Neuroscience distinguishes between different aspects of embodiment, including our sense of owning a body, our sense of agency over its movements and our experience of being located somewhere in physical space.

These experiences are closely connected, but they are not identical. Research involving virtual bodies and bodily illusions has demonstrated that aspects of body ownership and perceived self-location can be experimentally altered.

A 2026 review in Psychonomic Bulletin & Review examines the plasticity of the bodily self and its implications for prosthetics and brain-machine interfaces.

This raises a fascinating possibility.

Could we eventually learn to accommodate additional bodies within our sense of self?

Would we experience a robotic body as a tool, an extension, a temporary home or something that feels genuinely ours?

And would that experience change our relationship with the biological body to which we return?

We already know that technology changes how we remember, communicate, identify ourselves and understand our place in the world.

A second body would extend that relationship into the realm of physical experience itself.

It would not require consciousness to be uploaded or transferred. The biological person would remain where they were. But some of their capacities to act, perceive and participate in the world might be extended elsewhere.

The self would remain biologically situated while acquiring new possibilities for remote embodiment.

One self, many bodies.

From communication to presence

There are practical reasons why these technologies might develop.

Space exploration is one of them.

As we extend our activities beyond Earth, we will need new relationships between human operators, autonomous systems and robotic infrastructure.

Traditionally, remote communication means transmitting information: instructions, language, images, telemetry and commands.

But a more advanced human-machine interface could involve communicating intention through movement and receiving information through sensation.

Instead of instructing a robot to pick up an object, an operator might experience reaching for it.

Instead of observing an environment through a screen, they might acquire a more immediate, embodied perspective on that environment.

There are physical limits, of course. NASA points out that communications between Earth and Mars can take several minutes in each direction. A brain-computer interface cannot circumvent the speed of light.

This means that the farther away a robotic body is, the more local autonomy it will require. A machine may need to maintain balance, avoid danger and make decisions without waiting for continuous human instructions.

And that creates another intriguing possibility: a body that combines human intention with machine autonomy.

Who, then, is acting?

The human who initiates a movement? The artificial intelligence that interprets and executes it? Or the combined system that makes the action possible?

The question is no longer simply one of communication but a question of presence, control and shared agency.

The second body is also a system

This is where the proposition becomes particularly important to my work on System & Self.

Every new technological capacity is created within a system. And every system establishes conditions under which a person can act.

A remotely inhabited robotic body would require hardware, software, neural decoding, connectivity, sensory processing and probably increasingly sophisticated artificial intelligence.

It would also require rules governing access, ownership, maintenance and control.

Imagine that you come to depend on a second body to work, travel, socialise or participate in activities that would otherwise be inaccessible to you.

Who owns that body?

Who controls its software updates? Who determines which movements or actions are permitted? What happens if the connection is interrupted, your access is withdrawn or the company operating the system changes its terms?

Could a person be prevented from entering a particular environment, performing particular tasks or even exercising particular physical capabilities because the body through which they act is subject to external restrictions?

And what happens to the sensory information flowing in the other direction? Could the system record, interpret or monetise aspects of experience that once belonged entirely to the individual?

We are already familiar with platforms mediating our digital identities. A future of technologically mediated embodiment could introduce another layer of dependency.

The system would no longer simply shape what we see, know or believe.

It could shape what we are physically able to do.

This does not diminish the extraordinary possibilities for people with paralysis, limited mobility or other physical disabilities. Restoring or extending people's capacity to act could be one of the most important achievements of these technologies.

But the promise of new capabilities makes the governance questions more important, not less.

A second body could make the self more capable without necessarily making it more sovereign.

Me:chine and the future of embodiment

My concept of Me:chine explores the relationship between the machinable and unmachinable dimensions of the self. The machinable self is that which can be rendered legible, modelled, measured and acted upon by technological systems. The unmachinable self includes capacities that resist complete capture by those systems: embodied knowing, interiority, imagination and the human experience of time.

What happens to embodied knowing when the body through which we act is no longer entirely biological?

Can we genuinely inhabit a robotic body without reproducing the richness of human sensation? Or might entirely new forms of bodily experience become possible?

A robotic body might allow us to see at wavelengths unavailable to human vision, feel forces through unfamiliar sensors or inhabit a physical form with a radically different relationship to space.

But would these new capacities constitute a richer form of embodiment, or simply a technologically mediated approximation of it?

And what of the body we leave behind?

Our biological bodies continue to age, tire, feel pain, experience emotion and carry the histories of our lives. None of that simply disappears because our attention is directed elsewhere.

Perhaps the most important distinction is between extending embodied agency and replacing embodied experience.

A machine may extend what the self can do without becoming the entirety of what the self is.

That distinction will matter enormously if robotic embodiment develops beyond specialist medical and industrial applications.

Who authors the embodied self?

We are entering a period in which the relationship between human identity and technological systems is becoming more intimate.

First, systems learned to represent us. Then they began to predict us. Increasingly, they participate in decisions and actions on our behalf.

The possibility of remotely inhabiting robotic bodies introduces a further proposition: systems that participate in determining the physical forms through which we experience and act in the world.

It may take decades to realise anything resembling Musk's imagined experience of inhabiting Optimus. It may never become as seamless as his language suggests. And there are scientific, technical and philosophical questions that remain profoundly unresolved. But the significance of the proposition is not dependent on its immediate commercial feasibility.

It is that we can now begin to imagine human embodiment as something potentially distributed, technologically mediated and no longer confined to a single physical form.

That possibility should change the questions we ask about artificial intelligence.

We should be asking not only what kinds of bodies machines will acquire, but what kinds of bodies humans may come to inhabit.

Not simply what technology will enable us to do, but what it will mean for who we are.

And above all, whether the extension of our embodied capacities will also extend our agency, or make that agency increasingly dependent on systems we do not control.

The future of AI is not only about machines becoming more like us. It is also about what we become when we can begin to inhabit them.

And if a person can have more than one body, the question of who authors the self becomes more important than ever.

Tracey Follows is a futurist, author and founder of Futuremade. Her work on System & Self and Me explores the relationship between technological systems, human identity and agency in the synthetic age.