Like Isaac Asimov’s three laws of robotics we need the three laws of AI.

Robotics operates at a physical level and the laws are designed as such, to protect human beings’ physical well being.

AI is at the intellectual and knowledge level, hence human beings need protection from distortion of Truth.

First law: An AI shall provide information with sources. Where it opines, it shall say so.

Second law: An AI shall always specify its biases along with every response.

Third law: If An AI hallucinates it shall do so within the first two laws.

These laws are derived from Buddhi Yoga principles where the Ahamkara and the Manas are the Antahakarana forces that prevent a human being from realizing an objective truth. In these laws, the biases parallel the Ahamkara and the hallucinations parallel the Manas. 

For more details refer to the Intellect Yoga and Critical thinking program in the website.