BackDual-Use Capabilities Enable Malicious Use and Misuse of LLMs
Dual-Use Capabilities Enable Malicious Use and Misuse of LLMs
"Like all technologies, LLMs have the possibility for misuse by malicious actors. Malicious use of dual- use capabilities of AI is a recurring concern within literature (Brundage et al., 2018; Hendrycks et al., 2023; Mozes et al., 2023)"(p. 84)
Entity— Who or what caused the harm
Intent— Whether the harm was intentional or accidental
Timing— Whether the risk is pre- or post-deployment
Other risks from Anwar et al. (2024) (26)
Agentic LLMs Pose Novel Risks
7.2 AI possessing dangerous capabilitiesAI systemOtherPost-deployment
Multi-Agent Safety Is Not Assured by Single-Agent Safety
7.6 Multi-agent risksOtherOtherOther
Corporate power may impeded effective governance
6.1 Power centralization and unfair distribution of benefitsOtherUnintentionalOther
Jailbreaks and Prompt Injections Threaten Security of LLMs
2.2 AI system security vulnerabilities and attacksOtherOtherOther
Vulnerability to Poisoning and Backdoors
2.2 AI system security vulnerabilities and attacksHumanIntentionalPre-deployment
Vulnerability to Poisoning and Backdoors > Natural Language Underspecifies Goals
7.1 AI pursuing its own goals in conflict with human goals or valuesOtherUnintentionalPre-deployment