For organisations
Why you are here
Thank you for your quick responses and openness to our ideas. We have brought you together on this page so you can discover one another and the possibilities where our knowledge, experience and work intersect.
What we need to consider together
OpenAI, Anthropic and Google are developing agents that act through tools. NVIDIA and Microsoft are advancing controls over their actions and authority.
NVIDIA says an agent can be isolated in milliseconds for crossing a boundary. But who protects the person when an allowed action does not serve their interests?
Technology is part of a wider picture: who sets goals and rules, controls resources and data, benefits and bears the consequences.
We have mapped these connections to inform our conversations and reveal opportunities where our work intersects.
The story of one quantum computer → A fictional monologue: everyone pursues their own plans. Who sees the whole picture?
The questions bringing our worlds together
Whose goals does a system serve when the person's interests and an organisation's diverge?
How can we assess the combined consequences of decisions that are individually acceptable but harmful together?
How do we verify authority at the moment of action when circumstances have changed?
What does each party actually need to know without accessing the person's whole life?
Who can stop, contest and remedy an action — and who is accountable for its consequences?
How can we independently demonstrate real benefit while preserving privacy?
How do we account for other people's rights, children's rights, and social and ecological consequences?
How can independent systems interoperate without dependence on a single owner — including us?
What small shared scenario could demonstrate value for people and organisations without rebuilding the whole market?
Existing work we can build on: OECD AI Principles · IEEE Global AIS Flourishing · AP2 · OpenID CAEP · W3C Verifiable Credentials · Kantara ANCR · A2A · NIST AI RMF · UNICEF · MyData · Solid · UCP · NVIDIA Open Agent Safety Platform · Microsoft Security · OpenAI agents · Anthropic computer use · Google computer use.
Our shared strength as engineers and architects is seeing the connections between these questions, learning from accumulated experience and building an answer from the person's side.
What we can create together
Help shape the Human-Side Protocol. We invite experts, researchers and standards communities to contribute requirements, challenge the architecture and help define its specification, safeguards and evaluation. Your expertise can influence the next version of the protocol.
Explore the draft and take part →
We see how these areas can connect by starting with the person and their whole life. Together, we want to shape shared rules for interaction that anyone can use, verify and improve, preserving safety and control on the person's side.
We have already built the human-side foundation. Shaping a common protocol requires the knowledge and experience we bring together.
MCP, Linux and blockchain show us different ways to connect systems, develop shared infrastructure and independently verify information. Together, we can find the combination that works for the human side.
What if, instead of billions of repeated searches and development efforts, we could connect people with what the world already knows and can do — and find suitable solutions to their tasks in seconds?
We are developing a common language of skills and capabilities so people, AI systems, science and businesses can discover and use one another's capabilities across industries, countries and technologies. An open, shared pool could connect these descriptions with real offers and requests.
What we do next is in our hands.
In dialogue
IEEE Standards Association
Global · Engineering standardsPart of IEEE, the worldwide engineering community with more than 500,000 members in over 190 countries.
AI's impact on human wellbeing and how to evaluate technology from the person's side.
Linux Foundation
Global · Open AI infrastructure · AAIFA global home for open-source infrastructure. Its Agentic AI Foundation hosts projects including MCP, goose and AGENTS.md.
Connecting the human side with open AI-agent infrastructure while preserving personal control.
Kantara Initiative / ANCR
Global · Digital identity · Consent & trustAn international community developing digital identity and consent specifications, including ANCR consent receipts.
Connecting verifiable consent with permissions, delegated actions and their results on the person's side.
ADRA
Europe · AI, Data & RoboticsThe association representing the private side of the European Partnership on AI, Data and Robotics under Horizon Europe.
Human OS and its protective architecture within European research and industry.
NIIS
Nordic region · Digital governmentNordic Institute for Interoperability Solutions. Develops X-Road and Harmony eDelivery for exchange between independent systems.
How the person's side can interact with public infrastructure under each institution's existing permissions.
Mental Health Europe
Europe · Human rights & wellbeingA European network advancing mental health, social inclusion and the rights of people with psychosocial disabilities.
Human autonomy, protection and the boundaries of AI assistance in everyday life.
Fogarty Foundation
Australia · Education · Youth & AIA Western Australian foundation advancing education, school leadership and young people's participation in shaping AI policy through initiatives including Generation AI.
How Human OS could connect education with all areas of a young person's life, supporting learning, autonomy, wellbeing and responsible technology.
Technology Centre Prague
Czechia · European researchA national centre for EU research programmes and innovation, connecting organisations with European research networks.
Research consortia for Human OS, human protection and interaction with existing infrastructure.
Idiap Research Institute
Switzerland · AI research · Human-centred AIA Swiss research institute of national importance, advancing human-centred AI, trustworthy technology and participatory research.
How an independent Human OS could protect autonomy and people's interests, and evaluate AI influence across connected systems and everyday life.
CETIC
Belgium · Applied research · AI & cybersecurityA Belgian applied research centre founded by three universities, connecting digital research with industry through European and regional projects.
Initial discussions on potential scientific cooperation around Human OS, trustworthy AI and human autonomy across connected systems.
Link Digital
Australia · Global open dataCo-steward of CKAN, the open-source platform behind government, research and organisational data portals worldwide.
Open data, verifiable information and intersections with the Objective Observer Initiative.
Validis
United Kingdom · Financial infrastructureFinancial data infrastructure for audit, lending and analysis, with controlled access through APIs and MCP.
How financial infrastructure could meet the person's wider life through limited permissions and controlled requests.
Instituto da Transformação Digital
Brazil · Digital transformationA Brazilian institute connecting research, education and organisations around digital transformation and innovation.
Human OS research, responsible AI and institutional connections in Brazil and Latin America.
These conversations do not imply formal partnerships or endorsements.
One life. More value for everyone.
Imagine someone planning a new home. A bank can offer financing. An insurer can help plan protection. Other services can help with the move and everyday needs. The decision also affects time, sleep, health and the rest of that person's life.
Each organisation understands its part. Life GPS helps the person see how everything fits together.
A bank could respond to a financial need while Life GPS helps the person see its consequences for work, family and future plans. The bank receives a clear, authorised request and only the information needed to respond. Life GPS is designed to coordinate routine interactions across services, while the person and each organisation retain control over their decisions.
With the person's permission, there could also be a way to check what happened afterward. Did the service solve the problem? Did the expected costs match reality? Which assumptions need correcting?
Banks and insurers could learn whether their offers fit real needs. Data and analytics services could compare forecasts with actual outcomes. The person could see what worked and adjust the next decision. Each participant would receive only the feedback authorised for its role.
An opportunity to test predictions, improve existing services and reduce mismatched offers — in the situations people actually live. These possibilities would be evaluated in future pilots.
People live their lives. We work in the background to make those lives easier.
I respect what you build. I want its value to reach further, with benefits for your organisation and the people it serves. You and I are people too. Whatever our roles, that is who progress should serve.
A connection built around real needs
Placement cannot be bought. The person's needs and permissions shape the interaction. The personal core is being tested on iOS; external integrations and joint pilots are the next stage.
Bring your capabilities to the human side.
Help shape how existing products, services and infrastructure can connect with people's real needs through the Human-Side Protocol and future pilots.