An independent research institute · advising states and public institutions

Sovereign capability for the age of autonomous AI.

As AI shifts from assistant to actor, the foundations of public trust — who you are dealing with, where information came from, and who is accountable — can no longer be assumed. The selfdriven Institute helps states build verifiable digital trust into the fabric of how they identify citizens, govern institutions, and hold autonomous systems to account.

2027
Year autonomous agents are projected to act across critical public services at scale
100%
Of digital communications a state can cryptographically verify with the right trust layer
3-of-N
Witness threshold securing every identity event — no single point of control
Open
Built on open standards (KERI, ACDC, vLEI) — no vendor lock-in, no foreign dependency
The sovereign challenge

Public trust was built for a world where you could tell who you were dealing with.

That world is ending. Generative and agentic AI dissolve the cues citizens, officials, and institutions have always relied on to judge what is real and who is responsible. States that do not rebuild trust on verifiable foundations will govern in an environment they cannot authenticate.

Impersonation at scale

Synthetic voices, faces, and documents make impersonation of officials, agencies, and citizens trivial. "It looks legitimate" is no longer evidence of anything.

Provenance collapse

When any record, image, or statement can be fabricated, the chain of evidence behind public decisions, elections, and entitlements becomes contestable.

Accountability gaps

Autonomous agents already initiate actions on behalf of organisations. Without delegated, revocable authority, no one can say which human is answerable for what an AI did.

Our focus

Independent research and capability for the institutions of state.

We are not a vendor. We are a research institute that helps governments understand the threat, set the standards, and build sovereign trust infrastructure on open foundations they control.

01 / Research & foresight

Evidence for policy

Peer-informed papers, briefings, and scenario work on AI risk, digital trust, and verifiable identity — written for ministers, regulators, and senior officials who need clarity, not hype.

02 / Standards & frameworks

Trust architecture

Reference frameworks for sovereign digital identity, communications verification, and agent accountability — grounded in open standards a state can adopt without surrendering control.

03 / Advisory

Capability building

Direct advisory to departments and agencies: threat assessment, architecture review, and a roadmap from pilot to national deployment, with your people building the capability.

04 / Futures literacy

Preparing the population

Education programs that prepare citizens, students, and public servants to live and work in a world reshaped by autonomous AI — because trust infrastructure is only as strong as the people using it.

The sovereign trust stack

Trust you own, on standards no one can revoke.

Sovereignty means the state controls its own root of trust. The architecture we research and recommend is built on open identity standards — autonomous identifiers, witnessed key events, and verifiable credentials — so authority is cryptographic, portable, and never dependent on a single provider or foreign jurisdiction.

Each layer is independently verifiable. A citizen, an official, an agency, or an AI agent can prove who they are and what they are authorised to do — and that authority can be delegated, scoped, and revoked at any time.

L1

Autonomous identifiers AID

Self-certifying identity for people, entities, and agents — no central registry to compromise or capture.

L2

Witnessed key events KEL

Every identity change is append-only, witnessed, and tamper-evident. A 3-of-N threshold removes any single point of control.

L3

Verifiable credentials ACDC

Roles, entitlements, and attestations are issued as cryptographically chained credentials anchored to the identity layer.

L4

Legal entity identity vLEI

Agencies and institutions hold globally verifiable organisational identity — the basis for trusted cross-border and inter-agency dealings.

L5

Delegated agent authority dip

AI agents act only under scoped, time-limited authority delegated from a named human. Accountability is built in, not bolted on.

Programs of work

Eight areas where states are most exposed — and most able to act.

Our research and advisory is organised around eight areas of focus, each mapped to a policy domain a government must own as autonomous AI matures.

01

Direction

National AI-trust strategy, doctrine, and regulatory roadmap.

02

Engagement

Inter-agency, allied, and standards-body relationships.

03

Enablement

Public-service skills, futures literacy, and citizen education.

04

Protocols

Sovereign identity standards and verification infrastructure.

05

Sustainability

Funding, stewardship, and long-term operational viability.

06

Processes

Service delivery, incident response, and operational assurance.

07

Accountability

Audit, transparency, and answerability for AI-driven action.

08

Organisational

Structure, governance, and the human conductor model.

Research & resources

The evidence base.

Selected papers and resources written for decision-makers. The full library is on the resources page.