Creating compelling and engaging tabletop games is a challenging process that requires deep understanding of game mechanics and aesthetic considerations. To facilitate novice game designers without mechanics-creation and illustration expertise to prototype tabletop games, we created a four-week workshop for game design students to use Generative AI (GenAI) to create table-top games. We observed different use strategies, depending on their expectations of GenAI. In use, both the current text-to-text and text-to-image GenAI have been found to lack the ability to integrate unique ideas into the game structure, often resulting in misunderstanding. Fixing GenAI’s generated results is more complex than manually modifying them. Participants in the study tended to view the GenAI more as a reliable resource assistant rather than a true collaborative game co-designer. This work provides insights into how game designers may apply text and image-based GenAI tools for creating and testing physical components for games.