About this Microcredential
As artificial intelligence becomes embedded in critical infrastructure, business operations and everyday applications, securing AI-enabled systems is increasingly important. While AI offers significant benefits, it can also introduce new vulnerabilities and attack surfaces – from manipulated data and adversarial inputs to privacy leakage and attacks on generative AI systems. Understanding these risks and how to mitigate them is essential to the secure and responsible deployment of AI.
The Security of AI Practical Extension will extend your existing AI security skills with hands-on application in realistic scenarios. In a dedicated virtual environment, you’ll identify and test vulnerabilities in AI models and systems, assess model robustness and defensive controls, and explore security, privacy, fairness and explainability. With the advanced practical skills you’ll gain, you’ll be equipped to assess AI security risks and recommend appropriate technical controls, risk treatments and governance measures.
Key features
Attack, test and defend AI systems
Explore adversarial examples, data poisoning, model stealing, prompt injection and privacy leakage, then test defensive approaches.
Use contemporary AI security tools and frameworks
Work with Python, Jupyter, scikit-learn, TensorFlow or PyTorch, MITRE ATLAS, OWASP GenAI resources and AI security tools.
Assess AI security in real-world environments
Examine AI risks in critical infrastructure, autonomous systems, consumer applications, healthcare and financial services.
Turn AI security findings into effective controls
Identify weaknesses, evaluate defensive options and translate technical findings into security, risk and governance recommendations.
