Agentic AI

LLM-based autonomous agents are a developing area of AI focused on creating systems that can perceive, reason, and act autonomously using large language models (LLMs). These agents use planning, memory (sensory, short-term, and long-term), and tools to accomplish tasks. They are applied in fields like social science, natural science, and engineering. Evaluation includes human assessments and objective metrics. Challenges include safety, bias, robustness, and memory management, including writing, reading, and summarizing information. These agents aim to be more flexible and efficient than traditional AI systems.