Sensing consent
How do you consent to a network sensing you when you never connected to it? Design the privacy UX for rooms, streets and infrastructure that can perceive presence, movement and objects through radio signals.
HAAM / Research / 6G interaction design
The network is becoming an interface.
6G is not interesting because websites might load faster. It is interesting because connectivity, sensing, distributed compute and AI are beginning to behave like one system. That changes what interaction designers are designing.
personagentdevicesphysical environmentnetworkedge AIcloud
Interaction design moves inside the network.
Qualcomm describes 6G as a system built around three native capabilities: connectivity, sensing and distributed compute. AI workloads can move between device, edge and cloud; personal devices can cooperate; applications and networks can exchange context; radio infrastructure can also sense the physical world.
The designer is no longer specifying only what appears on a screen. Somebody has to decide what an agent may perceive, how it asks permission, which device has attention, where computation happens, what information follows you through a space and what happens when confidence or connectivity changes.
Before we standardise what the system can do, we should work out how a human is supposed to experience it.
First prototype
What can the room see?
Integrated sensing creates a new consent problem. A space may be able to infer presence or movement without a person explicitly pairing a device. The interface needs to make invisible perception legible and controllable.
Six things to prototype before 6G arrives.
We do not need commercial 6G hardware to start. Phones, wearables, UWB, Wi-Fi, Bluetooth, cameras, local models and cloud models are enough to prototype the interaction rules now.
Personal device constellation
Stop treating the phone, watch, glasses, earbuds, laptop and car as separate computers. Design them as one distributed body that decides which device should see, speak, remember or compute.
Device / edge / cloud
Make compute placement understandable. A person should be able to express an intent like ‘this conversation never leaves my devices’ and have the system reorganize itself around privacy, latency, power and reliability.
Experience intent
Applications may need to express what the experience actually needs, not merely ask for bandwidth. Urgency, continuity, privacy, spatial accuracy, energy priority and acceptable failure become interaction parameters.
The interface of a sensing city
If networks gain a live model of the physical world, the city itself becomes queryable. Where are people actually walking? Where is it quiet but not isolated? Where are cyclists and cars repeatedly surprising each other?
Agent etiquette for infrastructure
Agents will negotiate with buildings, vehicles, robots, shops and other agents. We need interaction protocols for permission, disclosure, memory, delegation and revocation in physical space.
The useful moment is before the standard is finished.
3GPP Release 20 is doing the 6G studies now. Release 21 is intended to begin normative 6G work. Commercial introduction is broadly expected around 2030, which makes 2026 unusually early for products but unusually good for interaction research.
2026Publish interaction concepts and working prototypes.
2027Collaborate with testbeds, researchers, operators and EU projects.
2028Move prototypes into pre-commercial network environments.
2029Design against emerging normative specifications and real device constraints.
2030+Ship products, standards, tools and services shaped by the early interaction work.
Starting points / August 2026
01Qualcomm, 6G Foundry: Rewriting the mobile playbook for the AI era, May 2026↗02Qualcomm, 6G device-to-data-center transformation, March 2026↗033GPP Release 20: 6G studies now, Release 21 normative work next↗04University of Oulu 6G Flagship, Human-centric Wireless Services↗05SNS JU, 6G4Society and the case for reconnecting HCI with technical development↗06SNS JU, €116M for 20 new 6G projects and further 2026-2027 funding↗
Design the human layer before it becomes an afterthought.
HAAM is exploring interaction primitives for AI-native, sensing-native networks: consent, device collaboration, compute placement, application intent and agent etiquette in physical space.
Explore this with HAAM →