The development of Darwin Monkey marks a milestone in artificial brain emulation, bringing Artificial Intelligence closer to the cognitive capabilities of living beings. However, this advance also raises profound ethical dilemmas about the consciousness, responsibility and the social impact of possible general AI.
Researchers at Zhejiang University in China have announced the development of a new supercomputer called Darwin Monkey, also known as Wukong, designed to mimic the brain function of a macaque.
It is a neuromorphic artificial intelligence system, meaning it mimics brain function not only at the software level, but also at the hardware level, using chips designed to function as real neural networks.
The system is composed of 960 Darwin 3 chips, the third generation of neural processing units developed by the aforementioned university. Wukong can simulate more than 2 billion artificial neurons and over 100 billion synapses, a scale that brings it close to the neuronal complexity of a macaque brain.
It uses an artificial intelligence model called DeepSeek, which enables it to perform advanced cognitive tasks such as logical reasoning, content generation, mathematical problem solving, visual and auditory recognition, learning, and language.
Despite its great power, Darwin Monkey would have been designed to be relatively energy efficient, which is key for neuromorphic systems of this scale.
The system offers the ability not only to simulate the macaque brain, but also to emulate the brains of animals with less neuronal complexity, such as mice and zebrafish. This opens the door to numerous experiments in comparative neuroscience without the need to use real organisms.
However, it is also a further step in the development of AI systems capable not only of learning but also of processing sensory information in a similar way to how the biological brain does.
Thus, neuromorphic AI paves the way for so-called general AI (GAI), a type of artificial intelligence that would be capable of performing any human cognitive task. This would involve not only specific tasks such as translating text, playing chess, or recognizing images, but also reasoning, learning, adapting, making decisions, and so on.
This type of advanced artificial intelligence poses profound ethical and social challenges. These include issues such as the possibility that it might develop some form of consciousness or capacity for suffering; the eventual need to recognize its rights and, consequently, treat it ethically as a living being; the attribution of responsibility for the decisions it makes; its potential application in the military or in mass surveillance programs; the challenges of its regulation and the limits of its use; as well as the risk of widening the technological gap between countries or causing massive job losses.
Addressing these dilemmas is particularly complex, given that technological innovation advances at a rapid pace and the questions it raises simultaneously span technical, philosophical, legal, and social dimensions.
Lucía Gómez Tatay
Bioethics Observatory – Institute of Life Sciences
Catholic University of Valencia
