Planning for risk and governance

You know, human behavior can be contradictory at times. The same survival impulse that draws us together also encourages us to look at people who are different from us as potential threats. Biases around gender, race, age, and other factors can creep into any AI system because these systems are all built, designed, and trained by humans who bring unconscious biases to their work. We can't promise to prevent all biases from creeping into our program, but careful review of learner segmentation, content sorting, activity recommendations, and performance analysis will help you eliminate or mitigate the risk of many potential issues at launch. But your plan must also include ongoing governance. And here's why. One of the most exciting and potentially terrifying features of generative AI systems is that they learn from each interaction and adapt their programming accordingly, without human intervention. This means that a system that starts out very closely aligned to your original vision…