My team is working on a game that is similar to reverse pictionary but uses generative AI images from one player’s prompt. There are a few questions that are necessary to answer through our prototypes.
- Are people interested in guessing the prompt for an image made by generative AI?
- This question is important to answer because it is the core premise of the game
- To answer this question, we could pre-generate cards with images made by various prompts and then have people guess them. If people are interested in this task, then our assumption that people would be interested in this is confirmed.
- My guess is that people are interested in this task. Reverse pictionary already exists, but the personalization aspect possible with generative AI is important.
- Is it possible for people to improve their skills of guessing prompts for images?
- Our readings so far have shown that people have the most fun when their brains are really stimulated through learning. Ideally, we want our game to facilitate that type of curiosity and skill-building.
- To answer this question, we should see if people can improve upon their prompts and guesses across different rounds of the game.
- My guess is that people will improve their guessing skills if they can determine how to get better at asking questions and picking up on key elements in the image.
- How can we scaffold the game so that people can learn from other people’s guesses?
- It is important that people can make progress towards guessing the prompt of the image so that they feel like it is not just a random guessing game
- We could experiment with different forms of guessing: from most details given to least details given. Whichever one results in most user satisfaction and achieving the game’s goals would be best.
- I think that people will like having small hints from other player’s questions so that they can get closer to the answer.
