Garry TanLead source
"Using just 3 million tokens of advice from ordinary people facing complex moral dilemmas brought malicious behavior down to 0%, with a 28-fold jump in efficiency across model generations"Read original on X
StoryFirst seen
Ordinary moral advice cut malicious AI behavior to zero, research finds
The result points to a cheap training method that improved efficiency 28-fold across model generations.
Archived Updated
Everything it touches
0 names 1 voices 1 source posts
The story’s neighborhood
NAMESVOICESPOSTS
Explore the graph, or choose a connection below.
Follow the evidence
The source posts
This brief comes from Tech Twitter’s edited headlines. The posts below are the sources grouped with the story; statements in them belong to their authors.
How the headline evolved
Ordinary moral advice cut malicious AI behavior to zero, research finds
Ordinary moral advice cut AI malicious behavior to zero, research finds
Ordinary moral advice cut AI malicious behavior to zero in research
Ordinary moral advice cut AI malicious behavior to zero in research
Research says ordinary moral advice cut AI malicious behavior to zero