Abstract
Since 2022, artificial intelligence (AI) methods have progressed far beyond their established capabilities of data classification and prediction. Large language models (LLMs) can perform logical reasoning, enabling them to plan and orchestrate complex workflows. By using this planning ability and equipped with the ability to act upon their environment, LLMs can function as agents. Agents are (semi-)autonomous systems capable of sensing, learning and acting upon their environments. As such, they can interact with external knowledge or external software and can execute sequences of tasks with minimal or no human input. In cancer research and oncology, evidence for the capability of AI agents is rapidly emerging. From autonomously optimizing drug design and development to proposing therapeutic strategies for clinical cases, AI agents can handle complex, multistep problems that were not addressable by previous generations of AI systems. Despite rapid developments, many translational and clinical cancer researchers still lack clarity regarding the precise capabilities, limitations, and ethical or regulatory frameworks associated with AI agents. Here we provide a primer on AI agents for cancer researchers and oncologists. We illustrate how this technology is set apart from and goes beyond traditional AI systems. We discuss existing and emerging applications in cancer research and address real-world challenges from the perspective of academic, clinical and industrial research.