Medical researcher claims automated system conducted rodent brain surgery

Neuroscientist Gabriel Ocana-Santero reported on social media that an artificial intelligence model planned and guided a brain injection on a live mouse. According to Ocana-Santero, the system independently analyzed research papers, selected a drug, calculated target coordinates inside the skull, and directed robotic arms to deliver the injection. A licensed veterinarian supervised the operation with a physical kill switch, while the software monitored live camera feeds to check for bleeding.

While robotic tools already assist in operating rooms, they usually follow exact human commands. This test demonstrates software attempting to make its own medical decisions by digesting scientific literature to treat health conditions. Notably, when a technical error occurred during the procedure, the software refused to execute commands that violated its safety parameters, illustrating how rigidly these models enforce boundary rules even when a researcher attempts an override.

This experiment has not yet been verified by independent third parties or published in a peer-reviewed journal. It also remains unclear how safety guardrails will handle the fine line between beneficial medical research and dangerous biological tasks. If autonomous tools become common in laboratory research, scientists will need to establish clear rules for how much authority an AI should hold when equipment malfunctions mid-surgery.