The Department of Defense (DoD) has released the Responsible Artificial Intelligence and Implementation Pathway. In this report, the DoD discusses the implementation of a Responsible AI Approach (RAI). The RAI defines guiding principles for the process of creating and the use of AI for combat and non-combat purposes. The primary objective of the RAI is the focus on the safe, efficient, and secure development and use of AI.
As AI becomes further integrated into national security, the DoD needs to find a balance between security and efficiency when it comes to the use and development of AI. If not treated appropriately, AI is a double-edged sword. Adversaries can create exploitations within the AI production lifecycle for future use or unintended behavior can result in harm. As a result, the DoD has constructed the RAI.
The RAI focuses on ensuring that AI in any stage of development or use must meet its ethical standard and foundational tenets. To ensure this, the Office of the Chief Digital and Artificial Intelligence Officer (CDAO) will be coordinating the implementation and guidance of policy with the assistance of the organizational entities within the DoD. While the RAI’s objective is to restrict irresponsible and unethical use of AI, it is not meant to be a barrier to AI implementation. There are measures within the RAI implementation meant to ensure flexibility and efficiency. RAI encourages decentralized execution by having RAI integrated into the agencies in a way that fits into their preexistent workflows and structures. Similar measures within the RAI foster innovation and implement AI in an efficient manner.