AI Agent Safety and Security Considerations

Jerry Huang, Ken Huang (), Krystal Jackson and Chris Hughes
Additional contact information
Jerry Huang: The University of Chicago
Ken Huang: DistributedApps.ai
Krystal Jackson: UC Berkeley, University of California
Chris Hughes: Aquia

Chapter Chapter 12 in Agentic AI, 2025, pp 369-407 from Springer

Abstract: Abstract This chapter discusses various aspects of AI agent safety and security, exploring potential vulnerabilities, mitigation strategies, and governance practices to ensure reliable and ethical AI systems. It discusses the risks posed by both accidental failures and deliberate attacks, highlighting the need for rigorous testing, redundancy mechanisms, and robust defenses against adversarial threats. Key considerations for inter-agent communication, embodied AI security, and ongoing governance are discussed. By addressing these challenges, the chapter aims to advance the safe deployment and management of AI agents.

Keywords: AI agent safety; AI security; Adversarial attacks; Governance practices; Redundancy mechanisms; Inter-agent communication; Embodied AI security; Formal verification; Mitigation strategies; Goal alignment; Motivation drift; Representation drift; Safe shutdown mechanisms; Agent governance (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:prochp:978-3-031-90026-6_12

Ordering information: This item can be ordered from
http://www.springer.com/9783031900266

DOI: 10.1007/978-3-031-90026-6_12

Access Statistics for this chapter

More chapters in Progress in IS from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().