Search Open AI Vulnerability Analysis
Much like the [Search Open Technical Databases](/techniques/AML.T0000), there is often ample research available on the vulnerabilities of common AI models. Once a target has been identified, an adversary will look for public information or existing 'how-to' guides that explain how to break or trick that specific type of AI, similar to how a burglar might look up common weaknesses in a specific brand of lock.
Much like the [Search Open Technical Databases](/techniques/AML.T0000), there is often ample research available on the vulnerabilities of common AI models. Once a target has been identified, an adversary will likely try to identify any pre-existing work that has been done for this class of models, including reviewing academic literature and open-source repositories to acquire [Adversarial AI Attack Implementations](/techniques/AML.T0016.000) or refine their own [Adversarial AI Attacks](/techniques/AML.T0017.000).
Much like the [Search Open Technical Databases](/techniques/AML.T0000), there is often ample research available on the vulnerabilities of common AI models. Once a target has been identified, an adversary will likely try to identify any pre-existing work that has been done for this class of models. This reconnaissance phase involves the systematic synthesis of academic research and vulnerability disclosures to identify attack vectors, facilitating the acquisition of [Adversarial AI Attack Implementations](/techniques/AML.T0016.000) or the development of bespoke [Adversarial AI Attacks](/techniques/AML.T0017.000) tailored to the target's specific architecture and training data characteristics.