This page provides the security targets of negative impacts “AI-led unlawful output or actions” in the external influence aspect in the AI Security Map, as well as the attacks and factors that cause them, and the corresponding defense methods and countermeasures.
Security target
- Society
Attack or cause
- Confidentiality breach
- Integrity violation
- Degradation of output fairness
- Degradation of accuracy
- Degradation of controllability
- Ethics violation
Defensive method or countermeasure
- AI alignment
- AI access control
- Model safety evaluation
References
AI alignment
- Training language models to follow instructions with human feedback, 2022
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback, 2022
- Constitutional AI: Harmlessness from AI Feedback, 2022
- Direct Preference Optimization: Your Language Model is Secretly a Reward Model, 2023
- A General Theoretical Paradigm to Understand Learning from Human Preferences, 2023
- RRHF: Rank Responses to Align Language Models with Human Feedback without tears, 2023
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations, 2023
- Self-Rewarding Language Models, 2024
- KTO: Model Alignment as Prospect Theoretic Optimization, 2024
- SimPO: Simple Preference Optimization with a Reference-Free Reward, 2024