Verification and evaluation methods
Methods for independently examining claims, capabilities, performance, behaviour and operating conditions.
JUST VERIFY TRUST LAB
Researching the methods and infrastructure for independent AI verification.
The Trust Lab is Just Verify's research and development environment for exploring future verification methods, evidence models and assurance infrastructure.
It investigates questions that are not yet sufficiently mature for established solutions and develops clearer foundations for testing, evidence and responsible organisational reliance.
THE ROLE OF THE TRUST LAB
A dedicated environment for developing future verification capability.
The Trust Lab examines how independent AI verification may need to evolve as systems become more capable, autonomous, interconnected and embedded across organisational and physical environments.
- 01Initiative
Sets and advances the wider verification agenda.
- 02Think Tank
Brings organisational experience and perspectives into the work.
- 03Trust Lab
Researches and develops future verification capability.
- 04Solutions
Applies sufficiently developed capability to practical requirements.
CURRENT RESEARCH AREAS
Six connected areas advancing future verification capability.
These areas provide a current research focus rather than permanent boundaries. They may evolve as AI systems, organisational requirements and verification methods change.
Evidence, provenance and accountability
How evidence is created, sourced, preserved, connected to claims and made useful for independent judgement.
Agent identity, authority and permissions
How identity, delegated authority, tools, access, approvals and accountability can be understood and verified.
Reliability, runtime assurance and continuous verification
How material behaviour, conditions and change can be monitored, escalated and independently re-examined.
Sovereignty, infrastructure and dependency
How providers, jurisdictions, facilities, data flows and technical dependencies affect resilience and control.
Emerging environments and responsible reliance
How verification may extend across physical, spatial, decentralised and machine-to-machine environments.
HOW RESEARCH DEVELOPS
From an unresolved question to useful verification capability.
The route is adaptable rather than mandatory. A question may begin at the stage most relevant to its maturity, available evidence and practical context.
- 01
Frame
Define the verification question, system boundary, organisational relevance and evidence gap.
- 02
Investigate
Review available research, technical evidence, standards, benchmarks and practical experience.
- 03
Develop
Create frameworks, taxonomies, evidence models or candidate verification methods.
- 04
Test
Examine assumptions through analysis, modelling, experimentation or focused collaborative work.
- 05
Translate
Determine what the findings can inform, what remains unresolved and which next step is appropriate.
POTENTIAL OUTPUTS
What Trust Lab work may produce.
Outputs depend on the question, maturity of the work and quality of the available evidence.
Research notes and evidence reviews
Conceptual frameworks and taxonomies
Evidence models and verification methods
Experimental findings and prototypes
Recommendations and transition candidates
LONG-TERM RESEARCH DIRECTIONS
Exploring environments future verification may need to address.
These are research directions, not mature services or established products. They examine how identity, authority, evidence, interoperability, infrastructure and accountability may need to be verified as AI moves into new operating environments.
Trust Lab research pathway
The Trust Lab explores future verification methods, evidence models and infrastructure for emerging AI systems and environments. It is research and development, not a fixed service pipeline.
Physical and embodied AI
Sensing, action, boundaries, override and accountability where AI affects physical environments.
Spatial and immersive systems
User, device, space, permissions, content and the integrity of contextual experience.
Blockchain and decentralised systems
Network identity, evidence, transactions and governance across decentralised environments.
Machine-to-machine ecosystems
Machine identity, authority, transactions and accountability across connected systems.
RESEARCH AND CONTRIBUTION
Contribution begins with a relevant question, capability or body of evidence.
The Trust Lab may work with suitable organisations, researchers and technical contributors where their expertise, systems, evidence or operating experience can strengthen a defined area of independent inquiry.
Research expertise and independent inquiry
Research methods, technical analysis, standards, governance, law, security and assurance expertise.
Systems, infrastructure and technical evidence
Relevant models, agents, platforms, facilities, data flows, dependencies and technical operating evidence.
Organisational questions and practical environments
Real questions involving adoption, behaviour, evidence, authority, oversight, risk and responsible reliance.
INDEPENDENCE BOUNDARY
DISCUSS A RESEARCH COLLABORATION Contribution may inform the research, but it does not imply endorsement, partnership or control over the Trust Lab's methods, findings or conclusions.
EXPLORE THE TRUST LAB
Help develop the future of independent AI verification.
Speak with Just Verify about a research question, evidence model, verification method, technical collaboration or future Trust Lab direction.