With the 2023 release of several Artificial Intelligence and Machine Learning (AI/ML) applications, there has been a heightened interest across all industrial sectors in using these new tools. AI/ML has the potential to automate many tasks that previously were only achievable manually through humans as well as unlock new capabilities by utilizing many large and different data sources. However, the use of AI/ML introduces new threats as evident in existing applications. As a safety-critical industry with long lifecycles and a slow rate of change, AI/ML should only be deployed with appropriate safeguards. While Standards Development Organizations (SDOs) are developing standards for implementing AI/ML with protections against unintentional errors, the current work is not considering protections against intentional errors and attacks on AI/ML. It is imperative that the relevant working groups immediately include cybersecurity considerations so that the first issue of the AI/ML certification standards may allow approval of AI/ML in aviation for a safe and secure implementation and deployment.
