AIG has proven to solve the problem of generating lots of items quickly. But can AIG methods generate high-quality, non-enemy items at scale? Can the methods specify how different the resulting items should look? Can methods provide subject matter experts with estimates on how many items can be generated from a single model, allowing for iterative AIG model development? The answer is ‘yes’. Advanced methods in AIG processes, with humans in-the-loop, have embraced methodogies that ‘say yes’ to all of these questions.

The discussion will cover a description of the innovative approaches to cognitive and item modeling that produce high quality items and provide insights into all of the ‘yes’ questions posed. New AIG item modeling methods, along with the application of these methodologies through caloborations, will be shared during our session. We’ve been at this a long time, alongside our partner subject matter experts, and will be sharing sesasoned insights to effectively implement AIG to produce high-quality, differentiated items to scale according to user-defined qualities.