Article 69JHS ChatGPT’s alter ego, Dan: users jailbreak AI program to get around ethical safeguards

ChatGPT’s alter ego, Dan: users jailbreak AI program to get around ethical safeguards

by
Josh Taylor
from Technology | The Guardian on (#69JHS)

Certain prompts make the chatbot take on an uncensored persona who is free of the usual content standards

People are figuring out ways to bypass ChatGPT's content moderation guardrails, discovering a simple text exchange can open up the AI program to make statements not normally allowed.

While ChatGPT can answer most questions put to it, there are content standards in place aimed at limiting the creation of text that promotes hate speech, violence, misinformation and instructions on how to do things that are against the law.

Continue reading...
External Content
Source RSS or Atom Feed
Feed Location http://feeds.theguardian.com/theguardian/technology/rss
Feed Title Technology | The Guardian
Feed Link https://www.theguardian.com/us/technology
Feed Copyright Guardian News and Media Limited or its affiliated companies. All rights reserved. 2024
Reply 0 comments