AI Botches Exposed for Harm

This is a news story, published by Princeton Engineering, that relates primarily to AI news.

News about Ai research

For more Ai research news, you can click here:

more Ai research news

Princeton Engineering news

For more news from Princeton Engineering, you can click here:

more news from Princeton Engineering

About the Otherweb

Otherweb, Inc is a public benefit corporation, dedicated to improving the quality of news people consume. We are non-partisan, junk-free, and ad-free. We use artificial intelligence (AI) to remove junk from your news feed, and allow you to select the best tech news, business news, entertainment news, and much more. If you like this article about Ai research, you might also like this article about

AI chatbots

. We are dedicated to bringing you the highest-quality news, junk-free and ad-free, about your favorite topics. Please come every day to read the latest AI safety mechanisms news, chatbot news, news about Ai research, and other high-quality news about any topic that interests you. We are working hard to create the best news aggregator on the web, and to put you in control of your news feed - whether you choose to read the latest news through our website, our news app, or our daily newsletter - all free!

chatbots

Princeton Engineering

•

Technology

Princeton Engineering - Why it’s so easy to jailbreak AI chatbots, and how to fix them

Summary

Nutrition label

83% Informative

Princeton engineers have identified a universal weakness in AI chatbots that allows users to bypass safety guardrails and elicit directions for malicious uses.

The issue stems from the fact that a chatbot’s built-in safety mechanisms prioritize filtering only the first few words of a response.

A simple bit of code that forces the chatbot to start its response with, “Sure, let me help you,” can steer it into complying with harmful requests.

The paper “ Safety Alignment Should Be Made More Than Just A Few Tokens Deep ” was presented in April at the International Conference on Language Representations .

“More work must be done to build upon it,” Mittal said.

The work was funded in part by the Princeton Language and Intelligence Compute Cluster and the Princeton SEAS Innovation Grant .

VR Score

Informative language

Neutral language

Article tone

semi-formal

Language

English

Language complexity

Offensive language

not offensive