Safety · Concept
Fragility of Value under Imperfect Alignment
Formal model showing imperfect proxy alignment plus strong optimization can deploy η-catastrophic value functions, motivating quantilizers.
Forecast horizon
Connections
Connections · 3
How this node ties into the rest of the map, and the evidence behind each link.
from · supports 3/10
Steering Vectors for Chain-of-Thought Faithfulness Both target monitorability/alignment reliability: steering aims to surface instrumental reasoning that optimization pressure may hide.
+3 growth to · informs 3/10
GPAI Undermining Public-Service AI Safety Governance (Policing) Value-fragility results motivate limiting optimization pressure; GPAI public-service paper shows legacy safety metrics fail under general optimizers.
+3 growth to · drives 3/10
Flourishing Value Theory (FVT) for Post-AGI Economics Value fragility under proxy optimization motivates richer non-price theories of societal value such as FVT.
+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 →