Engage
Research engagement, without overclaiming
Tyche Institute offers research-oriented workshops and limited pilots for teams that want AI-assisted workflows to leave better evidence: provenance, integrity hashes, timestamps, model/tool metadata, human review events, delegation chains, and explicit limitations.
These are research and engineering engagements. They are not legal advice, compliance certification, eIDAS trust services, or production assurance.
Workshop
AI Evidence Engineering for Engineers
A 2-hour or half-day workshop on designing AI workflows that can be inspected later: evidence maps, package sketches, tamper thinking, human oversight events, and implementation next steps.
- For AI, research, civic-tech, and education-tech teams.
- Remote or in-person.
Pilot
AI Agent Evidence Readiness Pilot
A 4–6 week pilot that turns one non-confidential AI-assisted workflow into a signed or hash-verifiable evidence package with a reference verifier and limitations memo.
- Evidence map, schema, sample bundle, verifier, tamper test.
- For non-identity, non-eIDAS workflows.
Good fit
- AI teams that need better evidence hygiene before formal audits.
- Research groups publishing AI-assisted artifacts.
- Civic-data teams that want reproducible public reports.
- Educational AI teams that want teacher-visible recommendation traces.
Boundaries
Tyche does not provide legal advice, certify EU AI Act compliance, prove that AI outputs are true, or operate as an eIDAS trust service provider. The services above are research and engineering engagements.
Work touching identity services, eIDAS trust-service implementation, border-control solutions, or public-sector identity systems is outside the default scope and requires separate review before any engagement.
Contact
For a research workshop or pilot conversation, write to [email protected].