Verified

Contribution

The 2021 UNESCO Recommendation is the first inter-governmental framework to name AI literacy and the verification of AI outputs as legitimate educational objectives, with member-state implementation guidance.

In our sims

Every advisor recommendation in a MyEdMentor sim carries a verifiability tag, the student must explicitly accept, modify, or reject it before the decision commits. The act itself is recorded.

Contribution

The OECD AI Principles establish 'human-centred values' and 'transparency and explainability' as binding norms, AI systems should explain their advice, and humans should retain decision authority.

In our sims

Advisor outputs include a 'why this advice' explainability layer. Students can challenge it before committing to a decision; their challenge is the data on which Verified competence is graded.

Contribution

AACSB's ongoing work on AI competencies in business education is the first accreditation-body acknowledgement that AI-aware decision-making is now a core graduate competency, not a niche specialism.

In our sims

MyEdMentor is structured so that every product implements AACSB's AI-Employability mechanics: oral defence, AI-use statement, credibility score, verify-the-AI prompt, badges.

Contribution

Halpern's empirical work establishes that critical thinking transfers across domains only when students are explicitly taught the transferable structure, not when they merely practise within a single domain.

In our sims

Our verify-the-AI mechanic is explicitly framed and named across every discipline, students learn it as a transferable skill, not a domain-specific habit.

Contribution

Russell argues that the central professional competence of the next two decades will be the capacity to evaluate and override AI advice, particularly when the AI is confident and wrong.

In our sims

The credibility score in MyEdMentor rewards defended-and-correct over confident-and-wrong, and explicitly penalises the latter, building the discrimination habit Russell argues for.

Contribution

A growing body of post-ChatGPT empirical work documents systematic over-trust of large language model outputs by learners, with limited self-correction even after exposure to errors. This work is the empirical bedrock for Verified pedagogy.

In our sims

Sims deliberately seed subtly-wrong advisor outputs at a known rate per run. Detecting them is part of the assessable Verified competency, not an Easter egg.