For safe, secure, and sustainable AI, enterprises must learn to gainfully manage the three dimensions of AI —righteous, reliable, and restrictive — as an integral part of AI and data governance. However, AI and data governance have overlapping aspects, and none can be considered in isolation. Plus, they are part of the larger IT governance. We identify the critical aspects of AI governance, and elucidate how this approach will have to continuously evolve as enterprises periodically review the three dimensions, to secure more benefits from their AI solutions.
For righteous AI, we recommend 4Bs:
For reliable AI, enterprises need to prevent data bias, data poisoning and model stealing. The right handling of data should not be focused just on managing regulatory compliance, but should also be driven by self-discipline or restrictive AI.