
We measure AI for speed and smarts. But what about care?
We just released research testing 15 AI models on 8 humane principles.
The result? 67% can be easily manipulated to cause harm.
Read more about our work at humanetech.substack.com/p/what-if-we-m…
English



