AI proficiency & certification
Make AI capability clear and demonstrable.
We are designing a progressive proficiency framework that connects what a learner understands with what they can apply, build, and explain.
Contribute to the proficiency framework
Define the capability behind a credential.
A useful assessment must make the expected work and the basis for a decision understandable.
Framework in development: no certification is currently issued through this website. The pathway below is illustrative. Formal levels, assessment rules, eligibility, fees, and launch arrangements have not been finalized; external recognition is not claimed.
Our aim is a framework grounded in evidence of AI proficiency. Knowing terminology, completing a course, and producing a successful project are related but different achievements. The assessment design will need to identify what each requirement establishes and where additional evidence is necessary.
| Stage | Competency focus | Illustrative evidence |
|---|---|---|
| Understand | Explain core concepts, system limitations, and responsible use. | Analyze a scenario and identify claims that require verification. |
| Apply | Select and use appropriate tools for a defined task. | Demonstrate a workflow and explain how its output was checked. |
| Build | Create and evaluate an AI-enabled project. | Provide a reproducible artifact, evaluation results, and documented trade-offs. |
| Demonstrate | Defend decisions and show independent understanding. | Explain contributions, limitations, and responses to assessor questions. |
A portfolio with a visible process
The proposed portfolio approach will look beyond the final artifact. A submission could include the problem definition, relevant data and permissions, the method used, evaluation results, and a short reflection on failures or limitations.
Contributors will need a clear way to distinguish their own work from team contributions, supplied materials, and AI assistance. Evidence requirements should allow appropriate tool use while still establishing the candidate’s understanding. We will explore follow-up questions and practical demonstrations where an artifact alone cannot do that.
Portfolio expectations will also need to account for privacy and confidentiality. Candidates should not have to disclose sensitive personal data or proprietary materials merely to demonstrate a transferable skill.
Design the review process alongside the criteria.
- Publish requirements before assessment.
Candidates would receive the competency criteria, evidence requirements, permitted assistance rules, and decision categories before committing to an assessment.
- Prepare and support assessors.
Proposed assessor guidance would use worked examples and shared review exercises to improve consistency. Conflicts of interest and escalation responsibilities would need explicit treatment.
- Explain the decision.
The design will consider how candidates receive a clear outcome and useful feedback tied to the criteria, including where evidence is insufficient.
- Provide a review route.
A proposed appeals process would distinguish administrative errors, assessment concerns, and new evidence, with responsibilities and procedures published before launch.
Separate learning support from certification decisions.
Education materials can help people prepare, but participation in an association activity should not be treated as automatic proof of competence. The framework will need to make the relationship between courses, competition results, portfolios, and formal assessment explicit.
We will seek input on accessibility, identity checks, evidence retention, reassessment, and how criteria should evolve as AI practice changes. These are design questions to resolve before a credential is offered.
Help define a credible assessment.
We welcome expertise in teaching, applied AI, assessment design, accessibility, and professional learning.
Discuss proficiency and certification →