Researchers have constructed a functional digital replica of a fruit fly brain containing approximately 125,000 neurons connected by roughly 50 million synapses. The structural copy was built neuron by neuron using connectomics data, then placed inside a simulated body. When sensory signals entered the model, neural activity propagated through the network as it would in a living insect.
The achievement represents a structural reproduction of biological architecture rather than a neural network trained to imitate fly behavior. This distinction means the system reproduces the architecture that produces behavior rather than approximating it. For neuroscience, this allows researchers to observe neural activity within a faithful replica, perform lesions, apply stimulation, and conduct manipulations to observe effects on the system.
The research team has stated ambitions to apply the same methodology to mouse brains. A mouse connectome would contain approximately 70 million neurons, representing a significant increase in complexity. If successful, a mouse brain running inside a simulated body would constitute a digital organism driven by faithfully reproduced biological neural architecture.
This development raises practical questions about where such a digital organism exists and whether a structural copy becomes a subject at some point. These questions are direct consequences of this research direction rather than abstract philosophical considerations.
The implications for artificial intelligence differ from current approaches. Contemporary AI systems are trained on data, optimized for output, and largely opaque. Whole-brain emulation functions differently: it is copied rather than trained, and its internal operations are transparent as a wiring diagram rather than a black box. An emulated brain learns like an animal rather than like a large language model, representing a different kind of intelligence with different associated risks.
Potential applications include drug testing on emulated brains as a replacement for animal testing and studying neurological disorders in controlled digital environments. The ethical calculus changes when the test subject is silicon rather than tissue, but only if the silicon is not conscious. The threshold for consciousness remains unknown.
Researchers describe their work in terms of structural copies and connectomics data rather than minds or souls. The trajectory extends from fruit fly to mouse, with potential future steps toward primate and human brains. Each step multiplies both complexity and stakes.
The fly brain currently operates within its virtual body, responding to simulated sensory signals and moving. The milestone demonstrates that the technology functions, while the ethical framework for addressing these developments remains incomplete. The real consequence is not the ability to emulate a fly brain but the impending necessity to decide whether to emulate a mouse brain and what that decision entails.
