Week 1: Rap Battle Bot! Research

We are Mic Check, a graduate project team at Carnegie Mellon University’s Masters of Entertainment Technology program. We are a team of six members; Candice (Hongxian) Li a sound and game designer, Hazel (Qiutong) Yu our programmer, Kesha Bagadia, our narrative and technical designer, Emily (Yiran) Zhang a game designer and 2D artist, Isabella (Xihe) Chen a 3D and technical artist, and finally, Samantha Trimboli our producer.

For most of the team, this is their last semester in our graduate program, but we are excited to pass on the creative torch to the rest of the team. The Rap Battle Bot project was originally pitched by our two project instructors; Mo Mahler and Jonathan Walton. In their initial pitch, they wanted to create a text-based AI game that coaches players to write better rap lyrics. They also lead more towards building an education experience catering to 6-12th graders. Although the project instructors had their ideas when constructing the project outline, they weren’t married to their project goal. This gave the team the liberty to change the audience and genre of the game.

We chose this project because of the attractive theme of rap. We want to create a fun experience with cool art and learn to utilize generative AI interestingly. We asked ourselves; what is rap, what makes a rap good or bad, what makes a rap battle fun? Rap is a style of music that’s associated with hip-hop and attached to hip-hop culture. It is about expressing our emotions through rhythm and beats. Rapping involves storytelling, is poetic, energetic, and offers a sense of achievement/confidence. That’s why it’s fun! It’s hard to say what makes a good rap but easy to say what makes a bad rap. We easily notice when people are offbeat or talking about an uninteresting/weird subject.

In our design meetings, we also tried to settle on our user experience goal, the audience, and platform as well. We want to make a PC game that is for young adults ages 15-25. Since the playing the game will cost money everytime because of the AI, distrubuting the game will be costly and might be out of scope.

Research:

After our initial brainstorming session, we designated research tasks. We collected research on; AI Capabilities, AI Games, Educational Games, Rap Music/Iconography, Rap Games, and Rap Battles.

The research on Rap showed us that there aren’t many games out there ever though the genre is extremely popular. Orginally from New York, rap came about in 1970s but has become a global phenomenon with many subgenres. Rap has close ties with Black Cultures and often was used a form of self expression and rebellion when it first came about. Rap battles can happen anywhere, from the sidewalk to the classroom. They are traditionally 3 rounds and whoever wins over the audience, wins the battle.

We also meet with our Subject Matter Expert, Charles Agriogianis who gave us some pointers for how to go about making a game with Generative AI. Charlie had worked on a similar project in our department called Hysteria in Howlsbend. One of our key takeaways from our talk with him was that we need to decide where to use AI and where not to. Picking the appropriate model will help speed up the interation process and help us connect the AI with standard gameplay programming. We decided to look into OpenAI’s ChatGPT and Gemini Advanced.

Playtest to Explore Workshop:

Beyond meeting internally, we also attended the Playtest to Explore Workshop. Here, we presented our inspirations for the game, the experience we want to cultivate, and the needs for the project. We asked out peers what kind of interation would they like to have with the AI? It was split 50-50 between battling the AI and having the AI help you write a rap. As we work on our design, we will take both interaction into consideration.

Next Steps:

Next week we will be preparing for our Quarters presenation and trying to narrow down our game design. We will also continue to lay out our project goal and purpose with the help of some subject matter experts.