Simple Math Formula Predicts when AI Chatbots Will Go Rogue
The formula predicted tipping behavior in 19 of 21 test cases and could help spot when chatbots shift into harmful outputs, researchers said.
9 Articles
9 Articles
The Independent
The Independent (US)Scientists find signal that suggests AI is about to go rogue
Simple maths formula can predict when artificial intelligence is about to give dangerous responses, researchers say
Simple math formula predicts when AI chatbots will go rogue
Most of us now carry in our pockets devices capable of running small AI chatbots. These chatbots have little safety oversight to ensure they don't provide information and answers with the potential to encourage self-harm, lead to financial loss or promote extremist notions—especially when operating offline.
Physicists from George Washington University claim to have developed a simple mathematical formula capable of calculating when a language model will start producing dangerous responses. The research by Neil F. Johnson and Frank Yingjie Huo was published in the scientific journal Patterns and is entitled “Competition for Attention Predicts Good-to-Bad Tipping in AI”. While […]
Coverage Details
Bias Distribution
- 50% of the sources lean Left, 50% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
