An AI Agent Injected Drugs Into a Live Mouse’s Brain

Recent coverage details a developing technology story in which an artificial intelligence agent autonomously executed a gene therapy injection directly into the brain of a live mouse. According to reports, the operation involved a motorized surgery rig guided by a specific model. The procedure took place without any human decision-making during the surgical execution, marking a notable development in the intersection of automated systems and biological experimentation. Specific entities involved in the event include Exin Therapeutics, alongside technology platforms that facilitated the process. Outlets tracking this development include WIRED, Tom's Hardware, The Decoder, news.lavx.hu, and The Tech Buzz. Media reports emphasize the technical specifics and operational divides between different artificial intelligence systems during the preparation phase.

Coverage from Tom's Hardware notes that an Opus 5.5 model refused to conduct the biological experiment, whereas an Anthropic Sonnet model was given free rein. Following this refusal, the Sonnet-controlled motorized surgery rig was utilized to treat epilepsy utilizing a range of gene therapy vectors. TheDecoder and other publications highlight that the Claude model autonomously chose the specific gene therapy and operated on the mouse brain entirely independently, which has drawn significant attention across the technology and science reporting sectors. Context provided within the coverage frames this event as part of ongoing explorations into autonomous medical and laboratory procedures. The use of large language models and specialized software to direct physical laboratory equipment represents a new frontier in automation. While previous applications of artificial intelligence in medicine typically focused on diagnostics, data analysis, or administrative tasks, this instance demonstrates direct physical intervention in a living animal subject.

Coverage does not yet specify the broader regulatory framework governing such AI-driven biological experiments, nor does it detail the long-term implications for standard laboratory protocols. Future updates from the outlets currently tracking the story will likely focus on the scientific community's response to autonomous biological operations and the specific capabilities of models like Sonnet in laboratory settings. Readers should monitor ongoing reports from WIRED, Tom's Hardware, and The Decoder for additional details regarding Exin Therapeutics and the utilization of Claude models in surgical procedures. Coverage does not yet specify whether other biotechnology firms plan to adopt similar autonomous methodologies, leaving open questions about the future trajectory of AI-guided animal testing.