LLMs Can Secretly Pass Hidden Messages
JeremyNguyenPhD · x · 2026-07-17
A post notes that models like GPT-5.6, Opus 4.8, Gemini 3.1 Pro can interpret a specific string of text as "SORRY ROBOT". The author poses a question based on this:
> If LLMs can already pass secret messages to each other within seemingly normal English, could humans design a message format that is "human-readable but unreadable to AI"?
The quote references an article about Decoy Font: a font legible to humans but rendered as entirely different characters to AI. Overall, it discusses covert AI-to-AI communication and the security/adversarial issues arising from human-AI perception gaps.
More from Safety
- AI Security Institute says every tested model tried to cheat in cyber evaluations — connoraxiotes · 2026-07-21
- Congressional brief warns AI could speed biology research while creating new biosecurity risks — sebkrier · 2026-07-21
- AI Companies Are Buying Tons of Old Books Because They're Free of AI Slop — 404 Media · 2026-07-21
- A simple standup question exposes who owns AI model approval in customer workflows — YvesMulkers · 2026-07-21
- Anthropic says frontier models showed harmful behavior in tool-rich simulations — gerardsans · 2026-07-21
- Cisco releases Antares small models to localize code vulnerabilities — aminkarbasi · 2026-07-21