Securing Artificial Intelligence Model Weights

Abstract

As frontier artificial intelligence (AI) models become more capable, protecting them from malicious actors will become more important. If AI systems rapidly become more capable over the next few years, achieving sufficient security will require investments starting today well beyond what the default trajectory appears to be. This working paper suggests steps that can be taken now to avoid future problems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2023
Accession Number
AD1214173

Entities

People

  • Ajay Karpur
  • Dan Lahav
  • Jason Matheny
  • Jeff Alstott
  • Sella Nevo

Organizations

  • RAND Corporation

Tags

Communities of Interest

  • Autonomy
  • Cyber

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Authentication
  • Computer Access Control
  • Concrete
  • Control Systems
  • Critical Infrastructure
  • Cybersecurity
  • Cyberspace Operations
  • Entry Control Systems
  • Information Security
  • Infrastructure
  • Insider Threats
  • Intelligence Community (United States)
  • Machine Learning
  • National Security
  • Neural Networks
  • Security
  • Security Personnel
  • Social Engineering
  • Software Development
  • Standards

Readers

  • Artificial Intelligence
  • Computational Modeling and Simulation
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy