CivAI Launches AI Voice Game to Demonstrate the Future of AI

CivAI, a nonprofit that works to build public understanding of artificial intelligence through live demonstrations of the technology, has launched a new AI voice game to educate the public on the future of AI.

"We Need to Talk" allows users to interact with AI characters who react and respond in real time. The game was created using voice technology from the major AI companies.

"We're trying to overcome the indifference many people have around AI," said Lucas Hansen, co-founder of CivAI, in a statement. "ChatGPT can be very stupid, so it's easy to miss the forest for the trees. Having an authentic voice conversation with an AI—one that expresses emotions, makes intelligent arguments, and seems to have its own wants and desires—that is an eye-opening experience. Our goal with this game is to give people a deeply personal understanding of how impactful AI will be."

To progress through each level of the game, users speak to an AI character and persuade it to change its mind on a subject. The game showcases the capabilities of current AI models to emulate human emotions. The AI characters laugh, cry, whisper, and yell in their efforts to push back on the user's arguments.

The game offers two main paths: Heart to Heart and The Orion Project. In the first path, users engage in a series of emotional and tense conversations with AI personalities, ranging from talking with a sick friend to reuniting with an estranged father. In the second path, users inhabit a world where a fictional AI company is on the verge of creating superhuman AI. The users navigate tough decision points that might arise along the way, such as an AI researcher grappling with whether to report an alarming discovery to company leadership.

"For the development of AI to be safe and successful, you need a very precise set of events: a lot of things need to go right, and a lot of different people need to make the right decisions," Hansen said. "In some cases, it's not even clear what the right decision is. It's a very fragile path. We're already seeing AI companies needing to make tough calls around safety, and the toughest decisions are yet to come. You need to put a lot of faith in these companies if you assume that they'll always make the right call."