Cities of Things started in 2018 as a research initiative and design lab at Delft University of Technology, initiated by professor Elisa Giaccardi and visiting professor Iskander Smit, and has grown into a foundation with an active research community across design, technology, and urban studies.
Key programs and collaborations include the Wijkbots program with Rotterdam University of Applied Sciences, the CollectConnect Hub exploration with Springtime and INFO, the Generative Things exhibition with ThingsCon, the Civic Protocol Economies research program with Amsterdam University of Applied Sciences, and international partnerships spanning design research, robotics, and urban policy.
The 2026 Manifesto for Immersive AI; drawing on 31 expert conversations, almost a decade of applied research, two speculative design scenarios, and thematic analysis; is the latest comprehensive synthesis of this emerging field.
In this document we share our current 2026 vision, and indicate where it is relevant. We illustrate this with a speculative story “The Grid”.
Cities of Things,
Iskander Smit
New possibilities for products, services, and environments where intelligence enters physical space.
AI is leaving the screen and entering the room. Buildings that learn how their residents move. Delivery carts that know the neighborhood. Energy systems that coordinate across a city block without being asked. Mesh networks where devices reason locally and act collectively. Physical AI adds situational awareness as the final crucial component to complete the digital transition.
This creates entirely new design possibilities. Products that combine intelligence with location. Services that adapt over time through genuine relationships with people. Environments where the interplay between human and thing becomes the core experience. The organizations that learn to design for these possibilities first will define the categories that follow.
This is the domain of immersive AI; the condition of living and working inside intelligence, rather than visiting it on a screen, or even addressing it via a voice command. This reality offers new affordances for designing products and services.
Our explorative research learnt us about what is relevant and what are considerations for design and policy making.
The impact of immersive AI
When AI becomes an environmental condition, distributed across things, spaces, and infrastructure, it opens up new design territory.
Things that sense and respond to context. Not smart devices reporting to a cloud, but things that reason where they are. A cleaning unit that notices something in the soil sensors and pauses to investigate. An energy hub that forms a position before the residents meeting. When edge AI gives things their own local intelligence, products become participants rather than instruments.
Services that grow through relationship. A delivery cart that learns your household rhythm over three years and adjusts without being told. A building system that earns trust through months of subtle coordination. Immersive AI enables services where depth is built through interaction, not imported from a dataset, and where that relational quality becomes the competitive advantage.
Environments that coordinate as assemblages. Not a single smart system, but dozens of things that share what they have learned and adapt together. The refrigerator talks to the composting system. The irrigation negotiates with the energy cooperative. New combinations emerge that no single product team imagined.
New infrastructure for community coordination. Mesh networks, shared protocols, neighborhood-level governance for autonomous things. The technical possibility already exists for communities to build their own intelligent infrastructure; cooperative, open-source, locally controlled.
These are not speculative claims. They are design possibilities grounded in seven years of research, prototyping, and expert consultation at Cities of Things, synthesized in the 2026 Manifesto for Immersive AI.
The relevancy of immersive AI
Immersive AI is not a single domain. It surfaces wherever intelligence meets physical space. Here are some entry points:
You are developing energy products or grid services, and wondering what happens when distributed systems start coordinating autonomously, forming positions before stakeholders convene. How do you design for community energy cooperation, not just individual optimization? What does a neighborhood-scale energy commons look like in practice?
You are working in mobility, logistics, or autonomous systems, and facing the gap between what the technology can do and what cities and communities will accept. How do you build trust over time? How do you navigate governance requirements before they become regulation? What can years of street-level prototyping with neighborhood robots teach your deployment strategy?
You are designing buildings, housing, or spatial environments, and sensing that the intelligence layer is about to change what architecture means. Buildings are becoming participants, not containers. How do you design for the interplay between spatial experience and distributed intelligence? What happens when a building’s systems form relationships with each other and with the people inside?
You are developing health or care services, and navigating the tension between ambient intelligence that supports and surveillance that controls. How do you design for dignity? Where is the line between a system that anticipates needs and one that removes autonomy?
You are shaping public space or civic infrastructure, and are aware that digital systems are arriving in the streetscape without the democratic processes that physical infrastructure requires. How do you create governance before deployment? What does citizen sovereignty over intelligent infrastructure actually look like?
You are a design team, innovation lab, or strategy department, and physical AI, agentic AI, or immersive AI is on your horizon but not yet in your practice. You need someone who has spent years mapping this territory and can help you navigate it with both ambition and critical awareness.
The Manifesto: ten considerations for getting it right
The new design possibilities of immersive AI are real, but so are the risks. Intelligence that enters the room can enrich experience or silently erode agency. It can strengthen communities or fragment them. The difference is in the design.
The 2026 Manifesto for Immersive AI provides ten considerations for designing responsibly and effectively in this space, developed through conversations with more than 30 leading experts across design research, urban policy, robotics, architecture, and technology strategy, combined with seven years of applied research.
Divided in five clusters: Legibility, Humble, Honesty, Honesty, and Openness
legibility interplay
1.
| Design Legibility of Agency – (LA) When a thing acts, people should be able to understand that it acted, why, and on whose behalf. |
| Provide Predictive Symmetry – (PS) When a system predicts a person, that person should have meaningful access to the nature and basis of the prediction. |
| Reveal Assemblage Transparency – (AT) When things coordinate with other things around a person, the fact and scope of that coordination should be perceptible. |
intelligence is humble
2.
| Keep Peripheral Presence – (PP) Agentic intelligence should operate at the edge of attention, not the center. |
| Emerge Relational Integrity – (RI) The quality of a human-thing relationship should emerge from interaction itself, not be predetermined by external optimization. |
design honesty
3.
| Appreciate Friction in Interactions – (FI) People must retain access to resistance, surprise, and unoptimized experience. Seamless interactions corrupt sensemaking. |
| Incorporate Temporal Honesty – (TH) Systems should not simulate relational depth they have not earned. |
support community
4.
| Leverage Democratic Precedence – (DP) Communities define the environment they want before agentic systems are introduced. |
| Ensure Distributed Sovereignty – (DS) Data, models, and infrastructure should be subject to collective governance. |
provide openness
5.
| Open up to Build Back – (BB) People must be able to introduce their own things, modify assemblages, and keep technological possibilities open. |
Possible activities
These activities are based on the latest research into the state of Cities of Things and the resulting manifesto, combined with nearly 10 years of exploratory and practical research into Cities of Things at TU Delft and an independent centre of expertise.
Inspiration and reflection
From workshops to ongoing advisory engagement. I bring the manifesto, the speculative scenarios, and the expert perspectives into your team’s process.
As provocation, as framework, and as a grounded counterweight to hype.
Manifesto-based design review
Your product, service, or policy initiative is assessed against the ten considerations. Where does it excel? Where are the blind spots? What risks are invisible from inside the project? A structured review that produces actionable recommendations, grounded in research rather than opinion.
Domain-specific deep dives
The manifesto research covers a broad landscape. I produce focused synthesis for your domain; extracting the insights, tensions, and opportunities most relevant to your challenge. The result is not a generic report but a tailored perspective informed by expert knowledge.
Speculative design for your challenges
Speculative storytelling to make the implications of immersive AI tangible and debatable.
The output is a strategic tool that makes visible the second-order effects that roadmaps and business cases miss.
This text is also available as PDF file.