Accession Number:

AD1122120

Title:

Briefing to the Air Force Scientific Advisory Board: T and E Contributions to Avoiding Unintended Behaviors in Autonomous Systems

Descriptive Note:

[Technical Report, Briefing Charts]

Corporate Author:

INSTITUTE FOR DEFENSE ANALYSES ALEXANDRIA VA

Personal Author(s):

• Wojtan, Heather M.
• Porter, Daniel J.

Report Date:

2020-02-01

Pagination or Media Count:

75

Abstract:

Testers rely on making valid inferences about a systems performance in untested situations in order to evaluate and certify the system. However, when systems are black boxesas is frequently the case for AI-enabled technologiesthese inferences are not possible. Avoiding unintended behaviors in these systems, however, requires us to make valid predictions about behavior. This means we must have models of system decision makingan understanding of what causally drives systems to make one decision over another.

Descriptors:

• test and evaluation
• artificial intelligence
• autonomous systems
• air force
• human-machine systems
• abstracts
• autonomy
• contracts
• detectors
• simulations
• department of defense
• supervisory control
• verification
• virginia
• machine learning
• perception
• training
• validation
• actuators
• bayesian networks

Subject Categories:

• Cybernetics

Distribution Statement:

[A, Approved For Public Release]