In their game, Northeastern students Jacob Ohlson and William Wennik want you to do your best chatbot impersonation.
People use artificial intelligence for tasks like writing essays, resumes, emails and even personal messages. Meanwhile, teachers, employers and editors are working to identify content created by AI.
Is it easy to spot text written by AI? Do you think you’d be good at it?
Now, there’s a fun new way to sharpen your AI-detecting skills. Two first-year students at Northeastern University have created a card game that combines a classic analog pastime with modern technology, all centered around spotting AI-generated content.
In their game, called Bot Buster, Jacob Ohlson and William Wennik want you to do your best chatbot impersonation. “Try to become AI and convince your friends that you are no longer human,” the game creators urge the players.
The game’s format is similar to Cards Against Humanity, where a judge presents a humorous or unexpected prompt to a group of players. Each player writes a response on whiteboards, attempting to sound like an AI chatbot, such as ChatGPT, DeepSeek or Gemini. However, only one player is actually using AI to generate their response and it’s up to the judge to figure out who it is.
If the judge correctly identifies the AI-generated response five times, he is declared the ultimate Bot Buster. But if someone imitates AI and successfully deceives the judge, they get a point.
“In a sense, it is like a reverse Turing Test,” Ohlson says, referencing the method proposed in 1949 by Alan Turing, an English mathematician and computer scientist, for evaluating a machine’s ability to think like a human.
Ohlson, a game design major, and Wennik, a game art and animation major, created the game in their Foundations of Game Design class under the guidance of professor Celia Pearce, an award-winning game designer and founder of the IndieCade festival.
“Card games are super popular right now,” Pearce says. “And they are getting increasingly interesting, clever and complex.”
Her class teaches future game designers essential core skills for creating all types of games, from tabletop games to live-action role-playing and video games.
During the course, students complete three projects: a live action role-playing game, an escape room built from budget-friendly supplies and an original tabletop game prototype.
Although many creative ideas emerge from the class, Pearce says she immediately recognized the potential of Bot Buster when the two friends brought the first rough iteration of their game.
“It wasn’t just me,” she says. “Everybody in the class was like, ‘This is pretty great.’”
The game has the important characteristics for a successful party game, Pearce says.
“It’s funny. It’s creative. It’s fast-paced. But it also has a very pointed social commentary that’s very timely,” she says.
Pearce encouraged Ohlson and Wennik to pursue publishing the game even before it was finished.
“They kept refining it and refining it, and it got better and better,” she says. “It isn’t unusual for there to be a really good game in the class. What’s unusual is for the students to keep running with it once the class is over.”
Between the fall and spring semester, the two friends rebranded the game, created a website featuring an AI chatbot named RONN, or Response Oriented Neural Network, based on ChatGPT, and caught the attention of a publisher interested in bringing Bot Buster to big-box stores.
However, before mass production, they must secure funding for the initial batch of game packs. To achieve this, they plan to launch a Kickstarter campaign.
“This semester is all about publishing this game,” the duo says.
Seeing Ohlson and Wennik committing to taking their prototype further and looking for a way to publish the game, Pearce says, feels gratifying to her as a game design teacher.
“They have that extra go-getter vibe,” she says. “They learned all the skills they needed in the class and then they continued to apply those skills after they got out.”
Each round players pick role cards that determine who will use AI and who must rely on their own wit. The judge then picks a prompt and a modifier.
Some of the lighthearted prompts, written by Ohlson and Wennik without the help of AI, include “What happens if I never shower again?” “How can I win a fight against Mike Tyson?” and “What do I do if my parachute won’t open?”
While testing the game, the creators found that AI often generates serious, factual answers, which can be a bit dull for a party setting. To add more fun, they decided that the judge would also select a modifier for all players — an adjective like “flamboyant,” “poetic,” “religious,” or “rude” — to influence the style of the responses.
Modifiers are an example of a simple tweak that made a big difference in the game experience, Pearce says.
“They also add to replayability, too,” she says. “You can put the same prompt in [and a different modifier] and get different answers every time from AI.”
She believes the game has a strong expansion potential, too.
“A lot of this style of games have extension packs … which is a great way to keep the revenue flowing,” she says. “So if the game is successful, they can make more [prompts].”