AI Threat Guides

Ten security best practices for LLM applications, mapped to OWASP LLM Top 10. Covers model layer, application layer, data layer, and agentic AI security—including a scannable implementation checklist.

A 4-step methodology for detecting AI bias and assessing fairness in AI systems—covering data audits, fairness criterion selection, disparate impact testing, and production monitoring. Includes tools comparison and the fairness impossibility theorem.

A 5-phase AI incident response framework covering detection, containment, investigation, remediation, and regulatory reporting—including EU AI Act Article 62 obligations and AIID submission guidance.

Step-by-step workflow for identifying adversarial inputs targeting AI systems, including input validation, transformation testing, behavioral monitoring, and response procedures for security and ML teams.

Step-by-step workflow for auditing AI systems for discriminatory outcomes, including fairness metric selection, disaggregated evaluation, data auditing, and regulatory compliance guidance.

Step-by-step workflow for identifying AI-generated phishing emails and messages. Quick-reference checklists for email authentication, behavioral indicators, automated analysis, and organizational response.

Step-by-step workflow for evaluating whether text was written by a human or generated by an AI system. Covers manual indicators, automated detection tools, stylometric analysis, and responsible decision-making.

Step-by-step workflow for identifying and responding to data poisoning attacks on AI training data, fine-tuning corpora, and RAG knowledge bases. Covers pre-training inspection, during-training monitoring, post-deployment detection, and remediation.

Step-by-step workflow for evaluating suspected deepfake video, audio, or images. Quick-reference checklists for visual inspection, audio analysis, provenance verification, and escalation guidance.

Step-by-step workflow for evaluating suspected AI-cloned voice audio. Quick-reference checklists for audio analysis, prosodic inspection, automated detection, out-of-band verification, and escalation guidance.

Six layered architectural controls for defending LLM applications against prompt injection. Implementation-ready checklist with code examples, OWASP mapping, and multi-tenant guidance.

A 7-step framework for protecting organizations against AI threats—covering threat surface identification, governance controls, technical hardening, red teaming, monitoring, and incident response.

AI red teaming is the adversarial evaluation of LLMs and agentic AI systems before deployment, testing for jailbreaks, prompt injection, harmful outputs, and bias. A 4-phase methodology with tools comparison.

Step-by-step workflow for securing AI model supply chains, including model provenance verification, dependency scanning, data source authentication, third-party tool security, and ongoing supply chain monitoring.

A phased checklist for safe AI deployment covering security configuration, governance sign-off, testing gates, and post-deployment monitoring. Tied to real incident types from the topaithreats database and EU AI Act Article 9 requirements.