Tokenless

Jailbreaking

Stealing Reasoning Traces from Proprietary LLM APIs — Ilia Shumailov & Alexander Panfilov

Stealing Reasoning Traces from Proprietary LLM APIs — Ilia Shumailov & Alexander Panfilov

A paper by Ilia Shumailov and Alexander Panfilov exposes a critical vulnerability in proprietary LLM APIs: encrypted reasoning traces, returned for conversation state management, can be extracted and replayed. This enables universal jailbreaking, privacy leaks of sensitive user data, and poisoning of AI agent traces. The study highlights significant implications for AI safety, model monitorability due to opaque internal reasoning, and even subtle forms of "distillation" where smaller models mimic frontier ones. The discussion covers architectural and system-level defenses, advocating for rigorous scientific inquiry in AI safety research.

The Promptware Kill Chain: How Prompt Injection Becomes AI Malware

The Promptware Kill Chain: How Prompt Injection Becomes AI Malware

Promptware introduces a new class of AI malware leveraging prompt injections to exploit architectural flaws in LLMs. This summary details The Promptware Kill Chain, covering stages from initial access and jailbreaking to persistence, lateral movement, and real-world impact. It emphasizes the critical need for a Zero Trust approach, treating AI agents as hostile runtimes to defend against these sophisticated AI-native threats.

915: How to Jailbreak LLMs (and How to Prevent It) — with Michelle Yi

915: How to Jailbreak LLMs (and How to Prevent It) — with Michelle Yi

Tech leader and investor Michelle Yi discusses the critical technical aspects of building trustworthy AI systems. She delves into adversarial attack and defense mechanisms, including red teaming, data poisoning, prompt stealing, and "slop squatting," and explores how advanced concepts like Constitutional AI and World Models can create safer, more reliable AI.