Discover AI Artifacts
This is when a hacker looks through a company's private files to find the hidden tools and data used to build their AI. Think of it like a burglar scouting a house to find the blueprints and the specific equipment used to build a vault, so they can figure out the best way to break in later.
The process of reconnaissance where an adversary probes internal systems to locate and catalog AI-related assets. By identifying components like container registries, model repositories, and training data management systems, the attacker gains the necessary intelligence to plan targeted exfiltration or disruption of the AI pipeline.
A reconnaissance technique where adversaries search private sources to identify and enumerate AI learning artifacts residing within a target environment. These artifacts encompass the software stack utilized for model training and deployment, data management infrastructure, container registries, software repositories, and model zoos. This discovery phase provides the technical metadata required to profile the AI ecosystem, facilitating the tailoring of subsequent exploitation, exfiltration, or sabotage operations.