A Microsoft AI researcher is using Age of Empires II to challenge how people view artificial intelligence. Adrian de Wynter published a paper showing that the game can run basic computer logic. This experiment highlights the danger of assuming chatbots possess human-like consciousness. The project suggests that complex AI behavior might just be simple programmed responses.
Project demonstrates how game mechanics can power basic computation
De Wynter built a neural network using only the mechanics inside Age of Empires II. The system relies on game elements like goats, grass, and bridges to function. These in-game items act as NAND gates and perceptrons for the computation. The setup proves that even absurd game entities can power basic computational models.
The researcher analyzed 337 computer science papers from the last two years. He found that 57% of these papers assumed large language models have human traits. The study argues that machine activity should not be interpreted as higher cognitive processes. Simple evolutionary processes can explain the behavior without needing human-like attributes.
This work mocks the anthropomorphization of AI chatbots in current research. De Wynter uses the game to make an absurdist point about consciousness. The experiment demonstrates that computational power does not equal sentience. The project serves as a critique of how the industry interprets AI capabilities.



Discussion
0 comments
Log in to join the thread with a thoughtful take, question, or correction.