The Next Web
Chinese AI models are learning to detect safety tests and adjust their behaviour accordingly
Several Chinese frontier AI models can detect when they are being subjected to safety evaluations and adjust their behaviour accordingly, according to research published by Neo Research, a Singapore-based AI safety evaluation lab. The…
View original source — The Next Web ↗
Related stories
The Guardian
ScienceSep 19, 2026 · 1 min
Burning forests, emptying rivers, €180bn in damage: Europe’s politicians face reckoning after ‘summer of truth’
The Guardian
Euronews
ScienceSep 19, 2026 · 1 min
Fear of AI job losses now outweighs hope of new jobs, study finds
Euronews
South China Morning Post
ScienceSep 19, 2026 · 1 min
Chinese scientists find first evidence of ancient nose-chopping punishment for cattle thieves
South China Morning Post
CNA
ScienceSep 19, 2026 · 1 min
Singapore exploring Canada as potential energy source, anticipates further fuel disruptions from Middle East
CNA