The article imparts knowledge on securing AI systems, focusing extensively on the defensive strategies and tools to mitigate potential risks. It commences by presenting a thorough analysis of security incidents, citing the significance of engineering tools to identify and tackle risks in such scenarios. Encouraging the application of the Failure Mode and Effect Analysis (FMEA) concept, the write-up asserts that it not only aids in identifying failing system components but also their causes and effects. The work of Microsoft and Harvard University on the failure modes in machine learning has been highlighted, stating its key focus on the reasons behind AI solution failures. Moving further, the document elaborates on the importance of defensive controls, preaching the need for organizations to acquaint themselves with potential threats and weaknesses. The necessity to uphold due diligence, especially in terms of compliance with international laws and regulations, is stressed upon. The topic of Adversarial Robustness is explored next, describing it as the process of building machine learning models with security, privacy, and regulatory principles. The article then emphasizes the substantial role of AI programs in determining countermeasures, considering varying deployment scenarios, budget constraints, and changes in vendor relationships, among many other factors.