Safety · Concept
Fundamental Insecurity of LLMs to Adversarial Attacks
ICML-argued claim that LLMs cannot be made fully secure against hacks due to a structural flaw in how they operate.
Forecast horizon
Pushes safety strategy toward containment, monitoring, and use-case limits over perfect hardening.
Connections
Connections · 2
How this node ties into the rest of the map, and the evidence behind each link.
to · drives 4/10
Industry-Wide Jailbreak Severity Scoring Framework If full hardening is impossible, standardized jailbreak severity scoring and incident response become central mitigations.
+4 growth from · tracked by 3/10
OpenAI MIT TR coverage links ongoing model break/jailbreak incidents, including OpenAI cases, to structural LLM insecurity claims.
+3 growthSignal sources
Signal sources
Dated facts from primary sources in this direction.
US evaluation centre Jun 2025
In June 2025 the US AI Safety Institute was renamed the Center for AI Standards and Innovation (CAISI), pivoting toward security, standards and adversary-model assessment.
NIST → Frontier safeguards May 2025
Anthropic activated its ASL-3 deployment and security standard with Claude Opus 4 on 22 May 2025 — the first real-world trigger of a responsible-scaling tier, focused on blocking bio-weapon uplift.
Anthropic → Cross-border testing 2025
The International Network of AI Safety Institutes (launched Nov 2024) ran a third joint testing exercise focused on agentic AI systems across cyber and fraud strands.
European Commission — AI Office →