Why AI and telecom hold the key to the future of human connection

For years, my working day has followed the sun. A design decision leaves my desk in Stockholm in the afternoon, gets picked up by a team in the Americas, and from there gets carried across to colleagues in India. As a team, we have delivered around the clock, across three continents, for longer than I have kept count. It works.

But I have also learned, over all those years and all those handovers, exactly where the tools stop and the people have to take over. The network can move the work. It cannot yet fully move the feeling of being in the same room.

That long-distance gap turns out to be wired into us.

Researchers at MIT recently found that social isolation triggers neural craving responses in the brain that are strikingly similar to the responses to physical hunger¹. When you lack human connection, your body literally thinks it is starving.

We have six billion internet users². We have video calls, instant messaging and social media platforms that know what you had for breakfast and who you voted for. And yet, loneliness is widespread. 

The telecom industry has been very good at solving reachability. You can send a message to anyone, anywhere, almost instantly. That problem is solved. The next problem is harder, and it is not about speed or bandwidth in any traditional sense.

The next chapter of telecom is about the fidelity of human feeling at a distance. Presence as a service. And if that sounds like marketing, bear with me. It is actually engineering.

What connection are we actually striving for?

Let me propose three layers of connection, because engineers like layers.

Layer 1: Reachability. I can send you a message. That problem is solved.

Layer 2: Understanding. I grasp what you mean despite language, culture, and context getting in the way. We are partially there. Machine translation works, and cultural context can be mapped. Sarcasm remains a challenge for both machines and a surprising number of humans.

Layer 3: Presence. I feel you here with me. We share a moment in time, not just information. This is the frontier.

But here is the point. Technology's job, if we are serious about human connection, is to collapse the gap between Layer 1 and Layer 3. To give distance the emotional bandwidth of proximity. To make a video call feel less like staring at a screen and more like being in a room together.

The technology stack of presence

If presence is the product, what does the stack look like? Three layers, because again, engineers.

Before we go further, I want to make a note on where we are. Some of what follows is being engineered and standardized right now, and some of it is genuinely over the horizon. I will flag which is which as we go, because the difference matters.

The network layer is largely a roadmap conversation: 5G-Advanced is being deployed today and 6G is in active standardization. The AI layer is a mix of shipping capability and near-term research. The interface layer, especially brain-computer interfaces (BCIs) at scale, is the most speculative part of the picture. Treat it as a direction of travel, not a product timeline.

The network layer

5G-Advanced and 6G are being designed as AI-native systems. These are not networks with AI bolted on. AI-native network design, edge compute and network slicing are on the roadmap of every operator and vendor today. 
What matters for presence: deterministic latency, edge compute, and network slicing for traffic that cannot tolerate jitter. Physics still applies. If you are trying to transmit haptic feedback or brain-computer interface signals in real time, a 50ms delay is not "degraded quality." It is a broken experience. The brain notices. Immersion dies.

 

The AI layer

Real-time translation is the obvious one, but translating words is the easy part. Translating cultural context and emotional register is harder. Sarcasm in Finnish does not map cleanly to sarcasm in Brazilian Portuguese. The AI layer needs to mediate meaning, not just vocabulary.

Then there is affective computing: machines sensing mood through biometrics. Imagine something like an Oura Ring that helps you "understand the emotional state of the people around you."³ That sounds invasive until you consider its potential to complement the subtle cues we already rely on and perhaps help us understand each other a little better.

AI in this stack is not a feature and not icing on the cake. It has to be embedded in the infrastructure itself. It’s a mediator between minds.

The brain-computer interface layer

If we can somehow create and process a digital representation of a thought, we can then transfer that and maybe recreate it in another head. That is not science fiction. It is an engineering problem with a timeline. Invasive implants exist in clinical settings, and consumer neural wearables are early; thought-level communication at scale is a direction, not a dated roadmap item.

The path runs from invasive implants (expensive, medical, niche) to wearable sensors (consumer, scalable, less precise but good enough for intent and emotion). We do not need to read every neuron. We need to read enough. Add haptics at distance. Touch as a communication channel. Spatial audio. Volumetric video.

All three layers need to work as a unified system – network, AI and interface. You cannot get presence with any one of them alone.

Hero photo by Ecliptic Graphic on Unsplash

New services on the horizon

The services below sit at different distances from today. Collaboration at distance is a here-and-now problem we are already improving. Biometric empathy builds on wearables that ship now. Shared immersive experiences are early but real. Direct neuro-communication is the furthest out, and the most speculative.

BCI and neuro-communication

Imagine one extremely artistic person and another with an extremely analytical brain. There is a semantic gap between them that words struggle to bridge. The artist thinks in texture and color. The analytical person thinks in structure and logic. Common “language” is normally difficult to find.

Now imagine a BCI bridge between them. Not reading their thoughts in some surveillance sense. But sharing the shape of an idea. The texture of a concept before it gets formed into words. They could create something together that neither could have made alone. This is connection beyond natural language.

Biometric empathy

Sharing physiological states to enable enhanced care. A nurse who can feel (in some mediated, bounded way) what a patient is experiencing. A partner who knows their spouse is having a bad day before the phone call. Not to surveil, but to understand. The line between the two is a design challenge.

Shared immersive experiences

Co-watching, co-travel, co-designing, co-presence in spaces that do not physically exist. When people from different backgrounds share the same virtual moment, something interesting happens. The diversity of perspectives becomes an asset rather than a barrier. The same principle applies to shared experiences. Homogeneity is comfortable. Diversity is productive; it enables combinatorial innovation.

Collaboration at distance

We are still having video calls. In 2026.

I have spent years handing work from Europe to India to the Americas. It works, but it works because of people and process discipline, not because of the tools.

I wrote once that "working on the same data sets, within the same environments through the same tools and on the same projects is way more important than physical closeness." I still believe that. But the tools need to catch up to the principle. AI-mediated co-creation across time zones, applied to design, music, construction, innovation and telecoms. Not video calls with better lighting.

Steering and guardrails

BCI as a thought-capture device

A BCI does not read "thoughts." It captures electrophysiological correlates of cognitive states. Whether through invasive microelectrode arrays or non-invasive wearables, the output is a high-dimensional signal: tens of thousands of neural firing data points per second, compressed into a digital representation.

This encoding goes through signal acquisition, feature extraction and decoding. The result is a digitized approximation of intent, emotion, or a thought.

Here is the problem. This payload is intimate by nature. It is not a message someone chose to compose. It is a reading of the cognitive state. A BCI signal may carry more than the user intended. Subconscious affect. Unspoken associations. Things you did not mean to share.

Thought-capture devices must implement on-device filtering and consent gates before any signal leaves the hardware. You need cryptographic control over what neural data gets transmitted.

Network criticality: transporting thought safely

Once a neural encoding leaves the device and enters the network, the requirements become categorically different from anything telecom has delivered before.

Integrity and confidentiality. A corrupted video frame causes a glitch. A corrupted neural payload could inject false cognitive content into the recipient's decoder. That is the equivalent of putting words in someone's mind. Integrity verification becomes a safety requirement. Neural data is the most sensitive class of personal information that exists.

Latency and determinism. Collaborative thought and synchronized emotion require sub-10ms end-to-end latency with jitter bounded to microseconds. Network slicing must guarantee dedicated, preemption-resistant paths for neural traffic. Unlike video streaming where buffering is annoying but survivable, neural synchronization requires predictive network behavior and near real-time network characteristics.

Data provenance. I have written about linked data and the problem of Chinese whispers in knowledge systems. The same principle applies here: neural payloads in transit must carry provenance metadata. Who generated it, when, through what device, with what consent parameters. This creates an auditable chain from thought-capture to delivery.

The intelligent fabric of human connection

None of this works as a collection of separate gadgets and tools. It works because the network itself becomes an intelligent fabric. It senses what is happening, decides what to do, and acts on it in real time, from one end to the other.

An intelligent fabric is not a fantasy. It is the next step for networks that already watch their own conditions, predict demand and degradations, and reconfigure themselves. Further evolution will allow the networks to transmit a feeling of presence through a fast and reliable connection for the interface, compute close enough to each person so the AI can keep up with the requirements, and a trustworthy path for the most sensitive traffic. The building blocks, slicing, differentiated connectivity, AI edge, network autonomy, are already here. Carrying human presence on top of them is where we are heading.

Some of this is close and some of it is still far off, but the direction is set by what we choose to build now. Telecom has always been about shrinking distance. The telegraph turned weeks into seconds. Mobile freed us from being tied to one place. What comes next closes a harder gap: the distance between hearing someone and actually feeling what they mean.

For decades we measured networks by speed, coverage, and cost per bit. That still matters. But it is no longer enough. The networks that define this era will be judged by a simpler question: did the person on the other end feel present? Did it feel like being in the same room? Did the distance quietly disappear?

That is a real shift in what a network is for. We are moving from carrying data to carrying understanding. From counting bits per second to caring about how much of a shared presence actually get established.

Reference

1. Tomova, L., Wang, K. L., Thompson, T., et al. (2020). Acute social isolation evokes midbrain craving responses similar to hunger. Nature Neuroscience, 23, 1597–1605.
2. https://www.statista.com/statistics/273018/number-of-internet-users-worldwide/
3. Tom Hale, CEO of Oura, in an interview for WIRED: https://www.wired.com/sponsored/story/four-predictions-for-the-future-of-human-connection-sparks

Further reading

• Physical AI needs a nervous system
• Five ways a network digital twin enables safe autonomy
• AI’s future will be defined by the intelligent fabric
• Technology and telecom trends 2026