Building Trustworthy AI at Scale: An Integrated XAI and MLOps Framework for VLT Innovation

Through a current co-op project, International Game Technology (IGT) and the University of New Brunswick are developing a powerful artificial intelligence (AI) model to predict the performance of Video Lottery Terminal (VLT) games. However, the model currently functions as a “black box,” making it difficult for business teams like game designers to trust its recommendations without understanding the reasoning behind them. Furthermore, the process to generate these valuable insights is manual, slow, and not scalable for day-to-day use. This project aims to solve these challenges by building a single, integrated software tool. This tool will provide a user-friendly dashboard that clearly explains why the AI predicts certain game features will be successful, bridging the trust gap for creative teams. In the background, the tool will be powered by a fully automated and reliable system that ensures these data-driven insights are generated efficiently and are always up-to-date. The expected benefit for IGT is the transformation of this collaborative AI project from a complex data science asset into a practical, trustworthy, and scalable business tool that will accelerate innovation and lead to smarter, data-informed game design decisions.

Faculty Supervisor:

Shadi Aljendi

Student:

Partner:

IGT

Discipline:

Computer science

Sector:

Arts, entertainment and recreation; Information and cultural industries

University:

University of New Brunswick

Program: