Chinese tech creator Xiaoningzi ran a seven-day self-experiment to see what happens when most deliberate thinking is handed over to AI. After the test period, the headline result came from an fMRI scan: total whole-brain activation pixels fell from 44,503 before the experiment to 20,918 after it. Her reasoning score changed only slightly, but the time needed to finish the same kind of test increased sharply.
Baseline checks before the AI-heavy routine
Before starting, Xiaoningzi underwent an MRI scan at a medical imaging center to confirm that her brain structure was healthy. The scan also showed that her prefrontal cortex, the area tied to higher-order cognition, was highly active under normal conditions. She then took the Raven’s Advanced Progressive Matrices test, or APM, which is used to measure logic, response, and fluid intelligence. Her initial result was 35 out of 36, with 97.2% accuracy and an estimated IQ of 143.
Work, decisions, and media intake were handed to AI
During the seven days, she pushed as much “slow thinking” as possible onto AI systems. That included topic selection, script outlines, business replies, and even code demos. The handoff extended beyond work. For ordinary choices such as where to eat on the weekend, she also followed AI recommendations instead of making the decision herself.
Her content diet changed as well. She watched more than four hours a day of AI slop or AI short-form animated videos, content described in the source as highly stimulating but often weak in logic. The structure of the experiment was simple: reduce active reasoning, cut back on self-directed judgment, and let AI handle tasks that would normally require sustained cognitive effort.
Post-test scans pointed to lower overall activation
At the end of the week, the imaging data showed a large drop in overall brain activity during comparable tasks. The source described it as the brain entering a kind of “power-saving mode.” The change was large enough to stand out on its own, with total activation pixels dropping from 44,503 to 20,918.
There was also a shift in how the prefrontal cortex was being used. Tasks that had previously relied mostly on vision and intuition appeared to recruit the prefrontal cortex after the experiment, as if higher-level planning functions were compensating for weaker overall processing efficiency. In the source, this was framed as a compensatory effect rather than a sign of stronger cognition.
Reasoning speed and working memory weakened
Her follow-up APM result placed her estimated IQ at 140, down slightly from 143. The more striking change was speed. The pre-test took 43 minutes, while the post-test took 57 minutes, meaning the same level of logical reasoning required 14 extra minutes. That slowdown was presented as the clearest sign of reduced mental sharpness during the experiment.
Working memory also deteriorated in an N-back test designed to simulate multitasking. Before the experiment, she was able to handle multiple streams of information with relative ease. After seven days of outsourcing thinking to AI, the source says she made frequent errors and experienced frustration and impatience. The article notes that the experiment may not meet strict scientific standards, but within this single-subject test, the results pointed in one direction: less active thinking coincided with weaker cognitive efficiency.