Mapping the ethics of generative AI: A comprehensive scoping review
A structured taxonomy of generative AI ethics issues gives APS governance teams a reference checklist against which to test the completeness of their own risk frameworks.
Key points
- A scoping review identifies 378 normative issues across 19 ethical topic areas for generative AI systems.
- The taxonomy covers governance, fairness, safety, transparency, and hallucinations - all live APS concerns.
- The review notes literature imbalances, including disproportionate focus on risks and unsubstantiated scenarios.
Summary
A scoping review published via the MIT AI Risk Repository maps 378 normative issues in generative AI ethics across 19 topic areas, including fairness, safety, hallucinations, privacy, governance, transparency, and cybercrime. The taxonomy is ranked by prevalence in academic literature and is available as an interactive online tool. The paper also critiques imbalances in how these issues are covered, noting overemphasis on certain risks and gaps in evidence for others. It is aimed at scholars, practitioners, and policymakers seeking a structured overview of the ethical landscape.
Implications for Australian agencies
- Consider APS AI governance teams could assess the 19-topic taxonomy against their existing risk registers or ethical frameworks to identify coverage gaps.
- Monitor Policy teams developing or updating responsible AI guidance may want to monitor the MIT AI Risk Repository for further synthesised frameworks added over time.
Implications are AI-generated. Starting points, not advice.
"Mapping the ethics of generative AI: A comprehensive scoping review" Source: MIT AI Risk Repository – Blog Published: 25 September 2024 URL: https://airisk.mit.edu/blog/mapping-the-ethics-of-generative-ai-a-comprehensive-scoping-review A scoping review published via the MIT AI Risk Repository maps 378 normative issues in generative AI ethics across 19 topic areas, including fairness, safety, hallucinations, privacy, governance, transparency, and cybercrime. The taxonomy is ranked by prevalence in academic literature and is available as an interactive online tool. The paper also critiques imbalances in how these issues are covered, noting overemphasis on certain risks and gaps in evidence for others. It is aimed at scholars, practitioners, and policymakers seeking a structured overview of the ethical landscape. Implications for Australian agencies: - [Consider] APS AI governance teams could assess the 19-topic taxonomy against their existing risk registers or ethical frameworks to identify coverage gaps. - [Monitor] Policy teams developing or updating responsible AI guidance may want to monitor the MIT AI Risk Repository for further synthesised frameworks added over time. Retrieved from SIMS, 18 May 2026.