Risk Academy · Programme outline
Ethics and Responsible AI in Risk
Fairness, accountability, transparency, safety and human oversight of AI used in risk decisions.نواتج التعلّم
- Identify who is affected by a risk model and name the harms it could cause.
- Test a model for subgroup differences in error and explain what the test does not show.
- Document a model and its limits in a model card and an impact statement.
- Design human oversight with clear decision rights and an escalation route.
Existing course sources: ai-emerging-technology-risk, model-risk-management, risk-data-governance.
New course outlines: Responsible AI for Risk Decisions.
Modules, labs and assessments
- Harms and stakeholders
Topics: affected groups, harm types, consent, power.
Lab: Map stakeholders of a synthetic risk model. Assessment: Short applied task. - Fairness and error
Topics: error by subgroup, measurement limits, trade-offs.
Lab: Test a synthetic model by subgroup. Assessment: Evaluation report. - Transparency and documentation
Topics: model cards, data statements, audit trails.
Lab: Write a model card. Assessment: Documentation review. - Oversight and accountability
Topics: decision rights, escalation, redress, frameworks.
Lab: Design an oversight process. Assessment: Oversight design memo.
Capstone and credential
A model card, a subgroup evaluation and an oversight design for a synthetic risk model.
Completion earns a completion record and an achievement badge through the NXS Passport. It records that the learner finished the programme's assessed work. It does not certify competence and is not an accreditation.
المصادر
- NIST (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1.
- OECD (2019, updated 2024). Recommendation of the Council on Artificial Intelligence, OECD/LEGAL/0449.
- UNESCO (2021). Recommendation on the Ethics of Artificial Intelligence.
- Mitchell, M. et al. (2019). Model Cards for Model Reporting. FAT* 2019.