Cost-Efficient Privacy Enforcement for Email-Generating LLM Agents under Strategic Information Requests

Summary

Researchers propose local contract-conditioned rewriting as an enforcement layer for LLM agents that draft and send emails on a user's behalf. The method builds an information-release contract for each outgoing draft, rewrites it with a local Gemma rewriter, and invokes a final PrivacyChecker selectively. Under normal prompts, leakage falls from 11.41% to 4.03% while UsableRate rises from 85.2% to 92.6%, and under strategic attack prompts checker-only cost amplification drops from 1.32x to 1.02x.

Related items

• InfoResearchers Used a Public OpenAI Model to Build a $250,000 Chrome Full-Chain Exploit in Six DaysSame vendor · The Hacker News
• InfoNvidia, worth $5.5tn, donates $250m to UK for AI safety researchSame vendor · The Guardian Technology
• MediumAI agents aren’t ‘going rogue.’ Frontier labs are skipping security basics.Same vendor · CSO Online
• InfoGoogle DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detectionSame vendor · DeepMind Safety Research
• Infollm-gemini 0.35Same vendor · Simon Willison's Weblog