The Fundamental Differences Between Biological Consciousness and Artificial Intelligence

Recent research by neuroscientists Borjan Milinkovic from the Paris-Saclay Institute of Neuroscience and Jaan Aru from the University of Tartu has highlighted the intrinsic differences between biological brains and digital artificial intelligence systems. Their theoretical framework, termed "biological computationalism," posits that consciousness arises from specific physical features that digital computers fundamentally lack. This challenges the prevailing notion of computational functionalism, which suggests that consciousness can emerge from any medium capable of processing information correctly.

Study Finds

Digital systems operate through a clear separation of memory, processing, and software, allowing programmers to write code independently of the underlying hardware. In contrast, biological brains function at the intersection of discrete and continuous domains. Neurons perform computations through continuous physical processes—such as ion flows and electric fields—rather than relying solely on binary states. This hybrid computation allows a single biological neuron to achieve computational feats comparable to multi-layer artificial neural networks, thanks to the interplay between continuous membrane potentials and discrete spiking events.

Study Finds

Energy scarcity plays a crucial role in the development of consciousness. The human brain, although only 2% of body mass, consumes approximately 20% of the body's total energy. To manage complex tasks efficiently, the brain reuses computational work across different scales, integrating continuous processes with discrete events. This "scale inseparability" means that molecular events influence network dynamics, creating a system where continuous and discrete processes are interdependent.

Study Finds

The findings suggest that current artificial intelligence technologies, including large language models and neuromorphic chips, may be limited by their reliance on discrete symbol manipulation. For artificial systems to achieve consciousness, they would need to meet three criteria: hybrid computation that combines continuous dynamics with discrete events, scale-inseparability shaped by energy constraints, and the ability to continuously modify their own physical structure. Emerging technologies, such as DishBrains—laboratory-grown neural cultures—and fluidic memristors, hint at alternatives that could potentially fulfill these requirements.

Study Finds

Some experts argue that the framework proposed by Milinkovic and Aru may overlook the potential for advanced algorithms to simulate aspects of consciousness. Critics suggest that improvements in computational techniques could eventually bridge the gap between biological and artificial systems, even if the current understanding of consciousness remains limited.

Study Finds

The research raises critical questions about the future of artificial intelligence and consciousness. While the theoretical framework emphasizes the need for radically different physical substrates, it remains uncertain whether such systems could genuinely support consciousness. The message for those pursuing conscious machines is clear: merely scaling up existing AI architectures is unlikely to suffice.

Study Finds

• “Brains operate at the interface of discrete and continuous domains,” — Borjan Milinkovic, Neuroscientist

Study Finds

• “The researchers call this “hybrid computation”—computation that is simultaneously continuous and discrete, where the algorithm cannot be separated from its physical implementation because the physics is the algorithm.” — Jaan Aru, Neuroscientist

Study Finds

• “Building artificial consciousness may depend less on better algorithms alone and more on radically different physical substrates capable of the continuous, scale-integrated, metabolically embedded processing that characterizes biological brains.” — Research Findings

Study Finds

This exploration into the nature of consciousness underscores the complexity of biological systems and the challenges faced in replicating such phenomena within artificial constructs.

Study Finds