.png?width=250&height=55&name=IA%20Brandmark%20green%20and%20Midnight%20green%20wordmark%20(1).png)
AI Due Process
A fair, documented process for any consequential decision AI contributes to.
Trust by Evidence
An AI Framework for Fair Assessment and Trusted Credentials
AI now touches identity verification, proctoring flags, authorship review, and credential validation. This whitepaper is a framework for making sure every one of those decisions can be explained, reviewed, appealed, and verified, before it's ever challenged.
ABOUT THE AUTHOR
The Governance Lens
Brandon Smith, CEO, Integrity Advocate
Brandon Smith leads Integrity Advocate's approach to assessment security, having spent over a decade in law enforcement before entering EdTech. Trust by Evidence extends that same instinct for accountable process into the governance of AI-assisted decisions, connecting due process, learner rights, and credential integrity into one working model.
THE PROBLEM
Adoption Isn't the Challenge Anymore
AI is already embedded in assessment. Identity verification, proctoring flags, authorship analysis, scoring, and fraud detection all increasingly run through AI. That part of the shift is largely settled.
Defensibility is what's unresolved. When an outcome gets challenged, an institution has to answer a specific set of questions: was there notice, meaningful human review, evidence, and an appeal path? If the answer is unclear, the outcome is not defensible, and the institution is exposed along with the learner.
Education already had due process before AI arrived. Grade appeals, exam reconsiderations, and academic integrity hearings are not new. AI does not remove that responsibility. It raises the stakes.
THE FRAMEWORK
Three Ideas, One Model for Defensible Outcomes
Trust by Evidence connects three ideas that are usually treated separately. Individually, each is familiar. Together, they hold up under scrutiny from the people affected by a decision and the institutions accountable for it.
Learner Rights
Notice, human review, explanation, evidence, appeal, proportionality, and verification.
Credential Security Trifecta
A secure chain of trust connecting learning, assessment, and credentialing.
This is what makes the framework durable. Regulations will keep evolving and AI tools will keep changing, but a model built around process, evidence, and accountable ownership holds up anyway.
AI DUE PROCESS
What Happens When a Learner Says "That's Not Fair"?
Adopting AI isn't enough on its own. Institutions need a documented, contestable path for every consequential decision AI contributes to, from proctoring flags to identity verification to authorship analysis. This layer protects learners from bias and protects institutions from legal risk with the same evidence-backed, time-bound review process.
An algorithm can tell you something looked unusual. It can't tell you whether it was a violation.
THE LEARNER RIGHTS LAYER
Seven Rights Every Learner Deserves in an AI-Assisted World
Responsible AI isn't only about better technology. It's about protecting the people whose futures may be shaped by it. These rights don't make assessments less secure, they make the decisions more transparent, defensible, and worthy of trust.
- 1. Right to Know
- 2. Right to Human Review
- 3. Right to Exaplanation
- 4. Right to Evidence
- 5. Right to Appeal
- 6. Right to Proportionality
- 7. Right to Verification
THE CREDENTIAL SECURITY TRIFECTA
A Weakness in Any Layer Undermines the Rest
Learning, assessment, and credentialing form a chain of trust. If learning is unverifiable, assessment is weak. If assessment is weak, the credential is vulnerable. The Trifecta makes trust portable, letting legitimate learners prove their achievement to employers, licensing bodies, and credential evaluators long after the exam ends.
THE BIG SHIFT
The Next Challenge Isn't AI Adoption. It's Defensible Outcomes.
The old model of integrity was reactive: find cheating, stop fraud, invalidate results. The new model has to be broader, protecting legitimate learners, verifying real achievement, and building audit-ready systems that hold up long after the decision is made.
SCHEDULE A DEMO
See How the Framework Applies to Your Program
Every program's flags, appeals, and credentialing setup look a little different. We'll walk through how Trust by Evidence applies to your specific context and how Integrity Advocate's human-reviewed model supports it in practice.