The future of AI is bigger than the model
AI/O begins where AI leaves the model and becomes part of a working system.
The shift
The first phase of AI was about models. The next is about systems.
The first strong phase of the AI market centred on models, analytics, machine learning, generative AI and cloud services. The next phase is about how AI is built into real systems.
AI is moving into the data companies already hold, into devices and sensors, into wireless networks, into local and distributed computing, into human decisions and into physical processes.
Once that happens, the performance of the model alone no longer decides the outcome. What decides it is whether the whole system works.
AI/O names that integration and system-level field. It does not promise a single technology. It assembles the capabilities needed to connect AI to the world around it.
The unsolved part
Companies rarely lack AI ideas.They lack the route from an experiment to a system.
Five transitions decide whether AI reaches production. Each one carries a question that has to be answered before the next step is worth taking.
Idea to system
How are purpose, data, architecture and ownership brought together?
Model to device
How are performance, energy use, latency and integration resolved?
Cloud to edge
Where does AI need to run, and what can be distributed?
Prototype to use
How is the system tested, validated and fitted to an existing process?
Research to business
How does a research result become a solution that can be deployed and scaled?
Two meanings of AI
A model can answer.A system has to work.AI can impress in a demo. Intelligence has to work in the world.
Artificial Intelligence describes models, algorithms, reasoning, prediction and generation. Advanced Intelligence describes the system level, where AI is connected to data, devices, networks and the physical world, and supports the people working with it.
Advanced Intelligence is not a claim to a higher scientific category. It is a strategic concept for AI that is built into a working, connected and validated system.
Why Oulu
Oulu connects intelligence.
AI/O is not a general AI brand with the name of a city added afterwards. It rests on a combination of strengths that already exist here.
Wireless technologies, 5G and 6G
Connectivity, distribution and real-time behaviour for intelligent systems.
Edge and distributed computing
AI close to where data is created, across the device, edge, network and cloud continuum.
Microelectronics, sensors and embedded systems
Intelligence built into physical devices and machines.
Autonomous machines, vehicles and robotics
Perception, decision and action in real conditions.
Health and industrial technology
Demanding applications where system-level validation is what counts.
Test and piloting environments
The means to verify AI as part of a complete system.
Research and industry working together
Shared problems, shared capabilities and routes from research into use.
This position rests on the combination, not on owning any single AI technology. It has to be shown through research, systems, test environments and company cases.
Technology
Advanced Intelligence is built across six layers.
The four research areas are not one of the layers. They cut across all of them.
Physical environment
Real-world environments
Factories, vehicles, health environments and public services
Devices, sensors and physical systems
Devices, sensors, embedded systems and machines
Compute
Local, embedded and edge intelligence
Local AI, embedded AI, edge AI and physical AI
Connectivity and distributed intelligence
5G, 6G, networks and distributed computing
Applications
Compute, data and AI platforms
Sandboxes, AI Garage, AIC Lab and university computing
Research, validation and ecosystem services
Research collaboration, foresight, funding and validation
Compute
The same use case can move from an experiment to sovereign capacity.
Work does not have to restart when it outgrows the environment it began in. The continuum is one path, not five separate ones.
Experiment and prototype
University computing
Devices and test networks
Sovereign capacity
National capacity
Usage data, results and new questions travel back to research.
How work moves
Work continues at defined decision points, or it stops.
As work moves towards deployment, its scope grows, more partners are involved and the requirements become stricter. Decision points are set along that path. At each one the work either continues with a wider scope or is stopped.
A stop is a result as well. Establishing that an approach does not carry is far cheaper at validation than after deployment, and the evidence usually points to a different route worth taking.
Evidence
Research is the input.Working systems are the output.
Research is the source.Real-world systems are the outcome.
Research produces the methods, algorithms, architectures, data, components, evaluation methods and expertise. AI/O connects them to company needs, devices, networks, test environments and real processes.
Test environments connect research with reality.
Laboratories, computing environments, simulations, test networks and physical environments show whether intelligence works as part of a complete system, and not only in a demonstration.
Boundaries
What AI/O is, and what it is not.
A concept is only useful if its edges are clear. These limits are part of the definition, not a disclaimer attached to it.
AI/O is
- A brand for the field. The research, innovation and market field for advanced intelligence.
- Systems-led. AI connected to data, devices, networks and the physical world, supporting the people who work with it.
- Research-driven. Credibility comes from research, testing and verified results.
- One umbrella. Local, edge, embedded, physical, distributed and real-time AI belong to the same field.
AI/O is not
- A general AI centre. Not every AI topic belongs under AI/O.
- A single testbed or service. AI/O is a wider concept and a way of working.
- A claim to a new scientific level of AI. Advanced Intelligence is a strategic concept, not a scientific rank.
- A finished organisational structure. Responsibilities, ownership and partnerships are agreed separately.
Get in touch
Start a conversation.
Tell us what you are trying to build. We help you find the right people, environment or funding route.
The services, research areas and environments behind this are described on the main page.