Delineation of Deployment Areas

AI has tremendous potential in solving scientific problems such as folding proteins. However, we should be more circumspect about its potential for solving moral and social problems. Part of the problem arises out of existing biases in the input data. The problem gets further compounded on account of the blackbox in machine learning algorithms because of which we are unable to understand how the decision was reached. The safer course of action would be to learn from other countries’ experience in this regard and rectify the defects before testing out any such pilot project.

This idea is important because the impact of algorithmic injustice is felt heavily by all concerned and particularly by the members of marginalised social groups.