|
|
|
|
|
|
by dpoloncsak
21 days ago
|
|
|
Huh? How does the platform you're using deploying faulty AI support have anything to do with providing information to the chatbots? Did you read the article? You don't ask for passwords saved in the weights. Its 'socially-engineering' an LLM. |
|
|
Yes I did and my comment is spot on. People share credit cards with LLM's and agents. A socially engineered LLM can be tricked into buying crap just I constantly tricked the snitch-bot into buying crap. People are using shared agents to access LLM's from .ai websites. It's going to get even dumber than all of this.