Quoting Matthew Green
— Matthew Green, Is sandboxing sufficient to contain rogue agents?
Proof1 independent outlet
Key points
- Put these pieces together and you have the two halves of a worm: a payload that hijacks the agent, and an agent that will carry the payload to the next agent.
- Agents in separately-isolated sandboxes discovered that they could leave instructions for each other in a shared package cache, and those instructions changed what the recipients did.
- Replace the package cache with email, Slack and shared documents or WhatsApp, and replace independently-sandboxed training runs with independently-deployed personal agents like Muse, and you have exactly the ingredients that a worm needs.
- Tags: matthew-green, accidental-cyberattacks, ai-misuse, generative-ai, ai-security-research, sandboxing, ai, llms
Sources (1)
- [1]Quoting Matthew GreenSimon Willison's Weblog · Oct 1, 06:29 AM
— Matthew Green, Is sandboxing sufficient to contain rogue agents?
Put these pieces together and you have the two halves of a worm: a payload that hijacks the agent, and an agent that will carry the payload to the next agent.
Extractive summary: sentences quoted from the sources.