Tag: agentic AI safety
Browse our exclusive articles!
OpenAI Technical Goals for Safe, Scalable AI
Discover OpenAI's technical goals focused on building safe, ethical, and scalable AI systems that benefit humanity and foster collaboration.
Key Concrete Problems in AI Safety Research
Explore key concrete problems in AI safety, focusing on robustness, interpretability, value alignment, and scalable solutions for secure AI systems.
Avoiding Faulty Reward Functions in Reinforcement Learning
Learn how to design effective reward functions in reinforcement learning to prevent failures and ensure AI agents behave as intended.
AI Learning Human Preferences for Safer Systems
Discover how AI algorithms infer human preferences to align with values and improve safety, developed with DeepMind's safety team expertise.
RL-Teacher: Enhancing AI with Human Feedback
Discover how RL-Teacher uses human feedback to improve AI training, ensuring safer, efficient, and more reliable reinforcement learning systems.