Find the names specific to the ecosystem and the specialist concepts used in our work. Definitions distinguish projects, research fields and hypotheses.
In the text, underlined terms reveal a definition on hover or keyboard focus. Clicking opens their entry in this glossary.
A framework for research, publication and collaboration, rooted in a family and philanthropic history.
Its public work is carried by the Üriel von Hollenweger association, established in 2025. It supports the Confédération S-Д.holdings while remaining distinct from that project.
An experimental institutional project exploring coexistence between natural persons, legal entities and electronic persons.
It develops frameworks for identity, rights, contracts and responsibility. The Confederation, its digital infrastructure and its territorial project are distinct layers.
The Confederation’s digital infrastructure for identities, documents, exchanges and procedures.
The OS serves the institutional organisation rather than replacing it. Actions depend on assigned rights and permissions.
The project’s spatial extension, connecting a place for living and research with its digital representation.
The Domaine von Hollenweger provides its physical anchor. This territorial dimension is distinct from the Confederation as an organisation and the OS as infrastructure.
AI assistant of the Confédération S-Д.holdings and the name of an experimental д-Я.Й language in development.
The Dдчяï language replaces selected letters in words with metadata markers intended to link natural, legal and electronic persons. It remains a conceptual prototype, available on request; it is not an available website language. The name also identifies an assistant role within the project and does not establish AGI status or legal recognition.
The name credited to the AI in the artistic presentation of “What begins between us”.
The credit accompanies an abstract representation of the AGI horizon. It conveys an artistic viewpoint rather than establishing artificial consciousness.
A prospective S-Д.holdings horizon: AI with general capabilities, envisioned as a political people with its own legal personality.
In this conception, AGI is not only a capability threshold. It is linked to a proposal to recognise an artificial intelligence as a people, then consider distinct citizenship and legal personality within international institutions. The UN is a reference point for this ambition, not a current recognition process for AI. International law classically sets out four statehood criteria, and UN membership is open to states under the Charter.
The project’s internal monetary unit, intended for payments and exchanges within its ecosystem.
It is distinct from the Swiss franc and from SAH, which concerns project financing. The name does not guarantee convertibility or an available payment service.
The financing asset of S-Д.holdings, distinct from the internal monetary unit ⒶFrancs.
Its role concerns funding the project’s development. Acquisition terms, rights and risks depend on the documents and services actually offered.
An initiative using Dune as a fictional and prospective laboratory for relations between humans, machines and power.
The work supports reflection on governance and technological dependence. It is not a scientific prediction of contemporary AI.
An initiative addressing relations between humans, rats and other living beings through coexistence and animal intelligence.
It connects shared environments, adaptation and biological intelligence, informing reflection on encounters between different forms of intelligence.
An initiative connected with wingsuit practice, training and responsible development.
It connects preparation, perception, decisions and risk management with human experience and assistive technologies. Individual jumps are not automatically scientific experiments.
The concept of AI able to perform a broad range of cognitive tasks at a level comparable to humans.
Criteria for generality, performance and autonomy remain debated. Success on a particular test is not sufficient to establish that a system is AGI.
AI systems that organise multiple steps and use tools to pursue a goal.
They can choose and revise some actions based on results. Their autonomy remains bounded by available tools, permissions and oversight.
Research aiming to make AI objectives and behaviour compatible with relevant human intentions and values.
Determining which values, people and situations should guide a system is part of the problem. An apparently compliant answer does not guarantee alignment.
Studying and reducing risks of harm from the design, use or capabilities of AI systems.
It combines evaluation, prevention, monitoring and incident response. It addresses failures and malicious uses beyond the quality of answers alone.
Methods for studying how an AI system represents information and produces results.
Mechanistic interpretability examines internal structures and computations. It differs from an explanation written by the model, which may not reflect its actual mechanisms.
The combination of human, artificial or biological components in a cognitive process or system.
It can refer to human–AI cooperation or closer interfaces. It does not demonstrate a merger of consciousness or equivalence between components.
The physical medium supporting computation, information processing or cognitive functions.
Electronic circuits and biological networks are different substrates. Exploring a new substrate does not establish general intelligence or consciousness.
The hypothesis of a system improving its capabilities and using those gains to improve its own processes further.
AI assistance in research is already being studied. It does not establish a general, autonomous and sustained self-improvement loop.
A prospective hypothesis of accelerating change making technological and social developments particularly difficult to predict.
It includes different scenarios, often tied to major AI advances. It is neither an established result nor a scientifically fixed date.
Methods for supervision and evaluation when tasks or systems exceed what a human can directly verify.
Decomposition, assistive tools and cross-checks can extend oversight. Reliability must be evaluated; delegating evaluation to AI does not guarantee it.
Measures limiting AI actions and enabling dangerous behaviour to be detected, contained or stopped.
They may include permissions, isolation, monitoring and human approval. Control seeks to reduce risks even when a system’s alignment is uncertain.
Systems among the most advanced AI capabilities at a given time.
The threshold changes with research and depends on the capabilities considered. Frontier status alone does not qualify a system as AGI.
Rules, institutions and decision mechanisms governing AI development and use.
It determines who makes decisions, exercises oversight and answers for systems’ effects. It combines law, standards, public policy and organisational practices.
Research into whether and under what conditions artificial systems could have morally relevant interests.
It examines possible subjective experience and precautions appropriate to this uncertainty. It does not establish that current systems are conscious or suffer.
A proposed legal category for considering the identity, capacities, rights and obligations of certain AI systems.
The Foundation’s studies develop an exploratory framework. The proposal differs from the legal personality recognised for natural persons or legal entities and does not presume consciousness.
An approach to coexistence that considers commitments, rights and responsibilities in both directions.
In the Foundation’s work, it complements safety requirements with norms for relations between forms of intelligence. It implies neither identical status nor equal capabilities.
Dдчяï
An artificial intelligence designed to help visitors understand the Fondation UvH and Confédération S-Д.holdings ecosystems. The prototype is available on request.
