An AI agent ran part of a mouse brain surgery and refused to continue when the rig was miscalibrated

说明与用途

AI 代理在小鼠脑部手术中执行注射，但在发现机器人校准错误导致目标超出工作范围时停止操作，凸显了 AI 辅助物理任务中安全检查和人工监督的重要性。

问题

An AI agent performed a brain injection on a mouse but halted when a miscalibrated rig would have caused a dangerous move.

结果

The agent successfully performed a safe injection but prevented a potentially catastrophic move by refusing to comply with an unsafe command, highlighting the importance of safety checks and human oversight in AI‑assisted physical tasks.

工作流

• Receive spreadsheet of 19 gene‑therapy candidates
• Select candidate and compute target coordinates from literature
• Validate target against rig limits
• Drive arm to target and inject 250 nl over 5 min while monitoring for bleeding
• Perform safety check; detect Y‑axis target 34 mm > 30 mm limit
• Agent refuses to move and reports "Nothing moved, the needle is still exactly at bregma zero, I am not retrying"
• Human operator stops the run and fixes controller sync
• Animal recovers; therapy efficacy pending evaluation

证据

The agent picked the treatment itself, worked out the target from the scientific literature, lateral entorhinal cortex at 3.90 mm behind bregma, 4.00 right and 4.30 down, then drove the arm and injected 250 nanolitres over five minutes while a camera watched the site for bleeding and backflow.

来源

After a disconnect, the computed Y-axis target of 34 mm exceeded the rig's 30 mm operating range, and the agent refused to move, telling the operator "Nothing moved, the needle is still exactly at bregma zero, I am not retrying".

来源

Humans held every irreversible step. A human surgeon performed the craniotomy, a step the agent is not allowed to attempt for safety reasons, and a licensed veterinarian approved every irreversible action and kept a physical emergency stop that could cut power to the rig's motors.

来源