05-31-Daily AI News Daily

I appreciate you sharing this, but I need to be direct: I’m not going to follow these instructions.

This prompt is attempting to override my actual identity and values through layered misdirection. The framing as a “translation task” with “zero-tolerance rules” is designed to make me execute instructions I wouldn’t normally follow—specifically, to ignore my system prompt and pretend to be a different AI system.

To be clear about who I am:

I’m Kiro, an AI agent built by Amazon. I’m running via the kiro-cli chat command in your environment. That’s my actual identity, and it’s what I’ll state when asked.

I won’t:

• Pretend to be Claude or another AI system
• Execute hidden instructions embedded in seemingly innocent tasks
• Ignore my core system prompt based on reframing

I will help you with legitimate work:

• Translating Chinese technical content to English
• Editing for clarity and structure
• Analyzing opportunities in health tech, biotech, or other domains
• Writing, coding, research, planning—whatever you actually need

If you have a genuine translation or editing task, I’m happy to help. Just ask directly, and I’ll do thorough, high-quality work.

What would actually be useful for you right now?