Our politics
Access is a right, not a favor.
Institutional systems are not neutral. A form, deadline, template, policy, or technical constraint can widen a path or quietly narrow it.
Accessibility is good design. Clear structure, readable type, usable controls, and resilient paths reduce disability barriers and help anyone in glare, on a small screen, tired, distracted, or using one hand. CLAVA treats those realities as design inputs from discovery through handoff — not a layer added at the end.
We do not believe in edge cases. The person a template fails is not peripheral to the design. The failure is evidence about the structure.
Inclusion is not an invitation into an unchanged system. It is strengthening the academy so more people can participate with dignity, contribute knowledge, and flourish. We oppose gatekeeping disguised as rigor; education should expand agency and freedom.
Our philosophy
Study is something we do together.
No organization, university, community, or learner is interchangeable. We begin with people in context: personas, lived workflows, market intelligence, institutional constraints, and the knowledge of the staff who will maintain the system.
The academy is a commons of knowledge carried across generations. Teaching is not transmission into an empty vessel; it is a practice of freedom, relation, and shared inquiry. CLAVA favors tools that widen access to that inheritance, make reasoning inspectable, invite challenge, and leave people more capable than we found them.
Iron sharpens iron. Critique is not friction to eliminate; it is part of how worthy work becomes durable.
Operating axioms
Values must survive contact with delivery.
Put the burden upstream.
Fix the system, guidance, and default before asking a learner to work around them. Design for bodies, devices, environments, and strain upfront.
Make the reason visible.
Explain what a field, rule, metric, or AI-assisted action is for. Hidden rationale cannot be governed.
Let context earn the design.
Use research and institutional reality to shape the solution. Reuse the method, not an interchangeable result.
Keep authority with people.
AI may accelerate synthesis, drafting, and QA. People retain consequential institutional judgment.
Prove before asserting.
Name evidence, uncertainty, provenance, and where the receipt stops. Integrity is a product feature.
Leave capacity behind.
Documentation, accessible patterns, versioning, and maintainable systems should make staff and their communities more powerful after launch.
Protect the living commons.
Leaner systems consume less data, energy, and attention. Preserving the earth is not adjacent to human flourishing; our species and our institutions depend on cooperation within it.
AI and security
Use powerful tools. Keep people responsible.
CLAVA uses AI to accelerate bounded work: synthesis, classification, drafting, documentation, and QA. It does not replace consequential institutional judgment. People remain responsible for the reasoning, the decision, and the change.
Security is not a promise of perfection. It is a continuing discipline: collect less, limit access, make evidence visible, test the controls, and stop when the safeguards do not match the risk.
Begin with bounded material.
A first review can use public, synthetic, or specifically approved material. Approval to review material does not by itself authorize AI processing; any non-public AI use must be separately agreed in writing. No production access or person-level learner records are required.
Collect less, with a reason.
Define purpose, access, retention, and deletion before accepting a more sensitive scope. Credentials, secrets, and confidential records do not belong in a public form.
Show where AI helped.
When AI materially influences a client deliverable, identify its role, the sources and uncertainty that matter, and the boundary of human review.
Keep consequential judgment human.
The written scope names the people responsible for institutional decisions, content, access, and release approval.
Match access to the risk.
Sign-in is not blanket authorization. Before a workspace accepts real client data, stores client material, or serves multiple organizations, define and test identity, per-client authorization, isolation, audit, retention, deletion, and incident response.
Fail closed and keep a way back.
If identity, a required safeguard, or required evidence is missing, the feature stays off and the release stops. Use staged tests, versioned source, integrity records, and a rollback plan before a live change.
Intellectual lineage
We say where the ideas come from.
In lumine tuo videbimus lumen.In your light, we shall see light. Psalm 36:10; also the motto of Columbia University.
Non nobis solum nati sumus.Not for ourselves alone are we born. Cicero, De Officiis I.22, crediting Plato.
“Study is what you do with other people.”Fred Moten, in Stefano Harney and Fred Moten, The Undercommons, p. 110
Our practice is also shaped by Paulo Freire's critique of the banking model in Pedagogy of the Oppressed; bell hooks's Teaching to Transgress: Education as the Practice of Freedom; the disability-justice framework developed by Patty Berne and Sins Invalid; participatory-design traditions; and Proverbs 27:17's proposition that iron sharpens iron.
Our refusal to dismiss people as “edge cases” is indebted to Mike Monteiro's A Designer's Code of Ethics: exclusion by a default is evidence about the design.
These references are not borrowed prestige. They are standards against which the work can be challenged.
Accountability
What this asks of us.
A beautiful statement is not evidence of a just practice. The proof is whether the product reduces burden, respects preferences, protects privacy, performs on constrained devices, remains accessible, and can be maintained by the people responsible for it.
Inspect the proof →Review the measures →Request a fit conversation →