Why are Artificial Intelligence systems biased?
The world wide web and the internet have often been touted as impartial appraisers of information — and therefore reliable sources on which to base informed, objective decision making. But problems have arisen with the use of artificial intelligence systems used to gather and assess information from the web and the internet. Often, what AI provides reflects many ingrained societal biases, writes Fulton Schools Professor Subbarao Kambhampati, a leader of major AI organizations and one of the experts who started the Conference on AI, Ethics and Society. Today, reining in such harmful biases is among the most urgent tasks in managing the risks of data-driven AI technology, Kambhampati says. One the positive side, many research institutions, corporations and governments are aware of the problem and appear willing to help solve it.