Artificial Intelligence (AI) and Machine Learning (ML) Acquisition and Policy Implications
Abstract
This white paper is a high-level survey of a set of both actual and potential acquisition and policy implications of the use of Artificial Intelligence (AI) and Machine Learning (ML) technologies. In this context, implications are known current effects, as well as possible future effects of the use of these technologies across a number of different identified domains where those effects become manifest. Some of these implications are primary effects that occur as a direct result of the application of the technology (e.g., the need to review the ethics used in autonomous decision-making by AI and ML), while others are secondary effects that occur as a result of a primary effect (e.g., the need to access data that will then be used to train supervised ML). In this context, acquisition implications are those effects which may require changes to the way defense acquisition is conducted, such as the way that AI and ML-based systems are validated by the acquisition PMO. Broader policy implications are those effects that may be related to defense acquisition, but which fall outside of acquisition as it is conducted today, such as those of data understandability. Successfully addressing these implications will require updating both acquisition and other policies to support the way development will need be done to build AI and ML systems. In this white paper, both the implications and ways of effectively addressing and managing them are discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 2021
- Accession Number
- AD1122292
Entities
People
- William Novak
Organizations
- Carnegie Mellon University