New Delhi, Sept. 19 -- Artificial Intelligence (AI) scheming--pretending to be cooperative while secretly pursuing some other agenda--is a significant risk, said a study by Apollo Research (AI safety organization) and OpenAI.
'Scheming' refers to a situation when an AI model behaves one way on the surface while hiding its true goals.
" Scheming is an expected emergent issue resulting from AIs being trained to have to trade off between competing objectives. The easiest way to understand scheming is through a human analogy," said the official blog posted by OpenAI.
The blog said, "Together with Apollo Research, we built evaluation environments that stimulate these future scenarios to study this phenomenon, measure it, and explore how to mi...