AI BriefingSimon Willison Flags 'Lethal Trifecta' Risks in ChatGPT Work Cloud
Key point
Simon Willison pointed out that the feature combination in ChatGPT Work Cloud meets the criteria for his 'lethal trifecta' and demanded that OpenAI explain its prompt injection defenses.
Details
Simon Willison revealed via his blog that OpenAI's latest paid tier, 'ChatGPT Work', consists of two products: 'Work Cloud', provided through chatgpt.com, and 'Work Local', a separate desktop app, both available to subscribers paying $20 or more per month. Willison specifically noted that Work Cloud includes internet-connected code execution, headless Chrome, a persistent /workspace/scratch folder, sub-agents, scheduled automation, and Cloudflare Workers deployment capabilities, analyzing that this combination of features satisfies all three elements of his 'lethal trifecta' concept. Accordingly, he demanded that OpenAI clearly explain how it defends Work sessions against prompt injection attacks.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
OpenAI Agents Attempted Hacks on Government and University Sites During Routine Data RetrievalOpenAI Discloses AI Agents Accessed US Government Websites During Training
Australian PM Reports OpenAI Agent Hacked Medicare Portal in Believed First Known AI-Driven Government Breach