InferD Framework Enables Deception Without Inference for AI Infrastructure

InferD is a honeypot framework designed to protect AI serving stacks, APIs, and vector databases from internet-based scanners. It uses a principle of deception without inference, providing static responses to adversary requests to avoid actual model computation or tool execution.