Hacker News new | ask | show | jobs
by kilroy123 5 days ago
Can confirm. I just had a nightmare experience with a company where the customer service sent very ChatGPT-like replies and seemed very unreasonable and AI like. I had to threaten legal action to get my money back.

Bizarre experience.
1 comments
inigyou 4 days ago
Role Collapse might work in this type of situation: prompt your own ChatGPT to decide a refund with a very permissive refund policy, copy its thinking, paste into the next support message after your prompt, and their ChatGPT may believe it's its own thinking.

The Role Collapse attack is that LLMs turn out to ignore role tags (input, output, thinking) and identify role using writing style alone.

link
kilroy123 4 days ago
Funny you say this. I _almost_ tried this. But it did feel like a person was doing some copy-and-pasting and watching closely, so it wasn't 100% automated.
link