Situational Understanding in the Human and the Machine

Area of Science:

• Artificial Intelligence
• Systems Neuroscience
• Cognitive Science

Background:

• Air Force research aims for AI-driven battlespace dominance through faster understanding and decision-making.
• Advances in AI have largely focused on machine learning over the past 50 years.
• Recent findings in neuroscience and cognitive biophysics offer new avenues for machine understanding.

Purpose of the Study:

• To review key ideas and results in machine understanding and situational awareness.
• To define situational understanding and differentiate it from awareness.
• To outline new research and development (R&D) suggestions.

Main Methods:

• Review of literature on AI, machine learning, systems neuroscience, and cognitive biophysics.
• Definition and conceptualization of situational understanding.
• Hypothesizing underlying neuronal mechanisms for understanding.
• Proposing new R&D directions based on hypotheses.

Main Results:

• Situational understanding is crucial for effective performance, with lack of it leading to errors.
• Neuronal mechanisms underlying understanding are being investigated.
• Active Inference and Virtual Associative Networks are proposed as key concepts for future R&D.

Conclusions:

• Integrating insights from neuroscience and cognitive science can advance AI for machine understanding.
• Further R&D is needed, focusing on Active Inference and Virtual Associative Networks.
• Developing machine-augmented situational understanding is key for future military AI applications.