Towards Safer Generative Language Models: A Survey on Safety Risks, Evaluations, and Improvements

MIT AI Risk Repository – Blog(Global) 18 Sep 2024 55

Provides a structured taxonomy of LLM safety risks that APS governance and risk teams can use when assessing generative AI deployments.

  • MIT AI Risk Repository summarises a survey identifying seven core safety risks in generative language models.
  • Risk categories include toxic content, hallucination, privacy leakage, and malicious use - directly relevant to APS AI governance frameworks.
  • Survey is from 2023 (arXiv:2302.09270); useful as a taxonomy reference but not cutting-edge given rapid field evolution.
  • Consider Governance and risk teams could assess whether this taxonomy aligns with or usefully supplements existing agency AI risk frameworks and assessment templates.
  • Monitor Policy teams may want to monitor the MIT AI Risk Repository more broadly as a curated source of risk frameworks relevant to responsible AI work.

Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.

View original source