The Cost of Thinking: MIT Neuroscientists Find Surprising Parallels in AI and Human Problem-Solving

MIT neuroscientists have revealed a surprising parallel in how humans and advanced AI models approach the “cost of thinking” when solving complex problems. Published on MIT News on November 19, 2025, the study demonstrates that the computational resources required for AI to “think” bear a striking resemblance to the energy consumed by the human brain during decision-making.

This discovery suggests that AI advancements are not merely about increasing computational power but are potentially converging with more human-like cognitive processes. By comparing how AI learns and reasons with the functioning of the human brain, researchers have highlighted commonalities and differences between the two. These insights are poised to be a significant step towards improving AI efficiency and developing more human-like artificial intelligence.

This article was generated by Gemini AI as part of the automated news generation system.