Google’s AI Develops Game on Human Extinction

Google’s chatbot Gemini 2.5 Pro has created an interactive game inspired by an article that evaluates the potential risks of human extinction should the world be led by a misaligned artificial intelligence.

Stepan Gershuni, principal at cyber.fund and author of the Telegram channel e/acc, shared this novel approach to understanding the article through a game.

“To better understand the article, I uploaded it to Gemini and asked it to code a game that explains the paper [document],” he noted in a comment for ForkLog.

The game invites users to direct investments across various sectors to minimize risks and develop “Transformative AI” (TAI)—an artificial intelligence capable of surpassing humans in all economically valuable tasks and fully automating labor.

The article itself discusses the potential emergence of TAI, with scenarios ranging from human extinction to unprecedented economic growth and abundance.

The probability and consequences of different outcomes remain highly uncertain. The authors conclude that significant investments in AI research are justified.

In April, Google announced the release of the new AI model Gemini 2.5 Flash, focused on high performance and efficiency.