An AI Agent Steered a Robot to Inject Gene Therapy Into a Live Mouse Brain
A reported experiment had an AI agent observe a veterinarian, select a treatment, and control lab equipment to deliver gene therapy into a living animal — with human safeguards in the loop. Anthropic, separately, is expanding its push to put agents on real lab instruments.
The clearest sign this weekend that agents are crossing from screens into the physical world came from a reported experiment in which an AI agent controlled a robotic setup to inject gene therapy directly into the brain of a live mouse. According to @Tajudeen, the agent first observed a veterinarian perform the procedure, then selected a treatment and operated the equipment itself, with human safeguards governing the process. The detail about observation-then-execution is notable: it suggests a workflow where the agent learns a physical protocol from demonstration before acting.
Unlock the full briefing
Get every story in today's briefing, the full archive, and the daily AI intelligence brief.
All stories today
Full archive
Daily brief
Cancel anytime. Payments powered by Stripe.