Safety · Concept
AI Agent Data Leakage in Non-Adversarial Enterprise Scenarios
A joint Singapore–Korea AI Safety Institute evaluation found that none of three tested LLM agents achieved fully safe execution across 12 realistic non-adversarial enterprise tasks.
Forecast horizon
Enterprise AI agent deployments require data-minimization and access-boundary controls as baseline safety requirements.
Connections
Connections · 4
How this node ties into the rest of the map, and the evidence behind each link.
to · informs 5/10
AI Customer Support Agent Exploitation (Web Agent Safety) Joint Singapore-Korea evaluation of agent data leakage in non-adversarial scenarios directly informs the broader trend of AI customer support agent exploitation.
+5 growth to · applies to 4/10
AI Customer Support Agent Exploitation (Web Agent Safety) The Singapore–Korea joint evaluation of non-adversarial agent data leakage directly instantiates the broader trend of AI customer support and enterprise agent safety risks.
+4 growth to · constrains 3/10
AI Agents Reshaping Knowledge Work Data leakage risks in non-adversarial enterprise AI agent deployments constrain safe adoption of AI agents in knowledge work.
+3 growth from · informs 3/10
UK AI Safety Institute UK AISI participated in the joint Singapore-Korea evaluation of AI agent data leakage in non-adversarial enterprise scenarios.
+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 →