What is the role of LLMs in AI agents?

• A. They directly control robotic actuators without planning.
• B. They provide the foundation for understanding and generating human language.
• C. They replace all decision making.
• D. They primarily store huge logs of data.

Answer: (B)

LLMs in AI agents act as the language brain: they interpret user input, infer intent, and generate coherent, natural-sounding responses. This language capability is what lets an agent understand requests, ask clarifying questions, and communicate plans or results back to humans. Because the core function is handling language—interpreting meaning and producing text—the description that they provide the foundation for understanding and generating human language is the best fit.

They don’t typically directly control actuators without planning, since acting in the physical world requires additional systems for perception, sensing, and action. They also don’t replace all decision making; agents combine language reasoning with planning, tools, and rules to decide what to do. And storing huge logs of data is a storage task, not the primary role of language models.