AI/ML is already used worldwide in programming and malware development; is it possible harness this capability to rapidly develop applications and scripts that aid cybers operators to perform offensive/defensive operations. Also, what legal and ethical considerations (human in the loop, bias, etc.) exist when integrating AI/ML into the development process. Finally, what process and technical controls should be implemented to maintain positive control over the application and effects generated by its use.
Use of AI/ML to aide cyber operators must be considered to assist them perform complicated tasks on demand and at scale in rapidly changing environments. An AI/ML solution with access to a wide variety of information would provide a more comprehensive assessment of the problem than a human and render a list of possible solutions and rationales that cyber operators could consider for action. For example, this AI/ML assisted weaponeering and apportionment could provide operators probabilities of success, alternate/supplemental actions, and recommended actions. It could also simplify or possibly automate capability pairing and battle damage assessment (BDA) actions through target identification based on stated objectives; physical, functional, and target system critical target nodes assessments, assess weapon effectiveness (Offensive) or vulnerability of critical infrastructure (defensive), and establish indicators of non-kinetic success or inflection points for defensive action.