The very definition of intelligence is at the heart of the human-AI debate. While the human brain, a product of millions of years of evolution, exhibits a remarkable plasticity, emotional depth, and contextual understanding, AI, particularly in its current narrow form, demonstrates superhuman capabilities in specific, data-intensive tasks, see Fig.1 [1].

As observed, 1) the way a pyramidal neuron reacts to an input is influenced by where that input hits on the dendrite. Inputs that come in close to the soma have a direct impact on the neuron's firing rate, while feedback inputs that land on the apical dendrites can influence burst firing (P(Burst)), 2) the firing rates of both presynaptic and postsynaptic neurons, along with P(Burst), play a crucial role in controlling plasticity, which includes Long-Term Potentiation (LTP) and Long-Term Depression (LTD), 3) the way cortical neurons integrate feedback and feedforward inputs might help tackle the credit assignment problem in hierarchical Artificial Neural Networks (ANNs), and 4) a diagram illustrating how neuromodulation can be incorporated into ANNs. On the left, an error signal from a network disturbance is transmitted through a global neuromodulatory effect.

In the middle, error signals are conveyed via node-specific neuromodulatory inputs. On the right, different neuromodulatory inputs could be involved in signaling various error functions.

The traditional yardstick of intelligence, often measured by logical reasoning and problem-solving, is being continuously challenged by the advent of machines that can master complex games and perform intricate calculations at speeds unattainable by the human mind [2].

This paper will explore this evolving understanding of intelligence, moving beyond a simplistic binary to a more nuanced appreciation of their complementary strengths.

The human brain and AI systems evaluate data in quite distinct ways. With its associative and pattern-based reasoning [3], the architecture of the brain is a highly parallel, dispersed network of neurons.

This biological foundation enables intuitive leaps, innovative problem-solving [4–13], and an amazing capacity to learn from sparse and incomplete data.

On the other hand, modern AI, mostly powered by Deep Learning (DL), is great at pattern recognition in massive databases, see Fig. 2 and Fig. 3 [14].

Though AI can identify relationships with remarkable precision, the capacity to comprehend the "Why" behind the data still presents a major obstacle. These systems, however, frequently lack the causal reasoning and common-sense understanding that define human cognition [4–13], [15].

A pivotal, and perhaps the most profound, distinction lies in the realm of subjective experience, or consciousness. As [16] famously articulated, the "Hard problem" of consciousness–the question of why and how we have subjective experiences–remains unsolved.

While AI can simulate emotions and even generate seemingly introspective text, there is no evidence to suggest that these systems possess genuine phenomenal consciousness [17]. The integrated information theory proposed by [18] offers a framework for quantifying consciousness, a metric that current AI systems are far from satisfying.

Feinberg and Mallat [19] further illuminates the neural correlates of consciousness in the human brain, highlighting the intricate biological mechanisms that give rise to our inner world, a complexity yet to be replicated in silicon.

Creating Artificial General Intelligence (AGI)—an AI capable of comprehending, learning, and using its intelligence to address a wide range of issues, very much like a human being—is riddled with difficulties. Emphasizing the necessity of systems able to integrate many cognitive talents, Summerfield [20] describes the complex nature of this project. One of the major problems is the brittleness of present AI: Systems taught on one job frequently fail dramatically when confronted with a barely different one [4–13]. The creation of strong and ethically aligned AI presents another major difficulty, one [21] nicely brought up in their examination of the possible existential risks of superintelligence.

Wang et al. [22] promoted an approach of "Intelligence without representation" that stresses embodied cognition and real-world engagement, therefore providing a possible but difficult way ahead.

Instead of seeing the expansion of AI as a zero-sum game, the most hopeful future is in the synergy of human-AI cooperation. The individual talents of artificial and human intelligence are quite complimentary rather than mutually exclusive. AI can be a strong instrument to enhance human intelligence by processing enormous volumes of data and finding patterns that may go unnoticed by human perception [23].

This collaboration can help to speed up scientific research, improve medical diagnoses, and inspire fresh kinds of creative expression. Going forward, the attention must change from a competitive story to one of co-evolution, whereby the individual skills of both human and artificial minds are utilized to meet the difficult problems confronting mankind.

The debate over whether the human brain or AI is "smarter" is, in many respects, a bit misguided. It assumes there's a single definition of intelligence, which overlooks the incredible variety of cognitive skills we possess. The human brain, with its ability to think, create, and understand complex ideas, is truly a wonder of nature. On the other hand, AI brings its own strengths, showcasing immense processing power and analytical skills. The real goal isn't to build a machine that can perfectly mimic human thought, but rather to cultivate a partnership with our intelligent creations.

This collaboration could lead us into a new age of remarkable intellectual discovery and achievement. The path to understanding and utilizing intelligence in all its forms is an ongoing journey filled with dialogue and exploration.

This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

All data are included in the text.

The authors declare no conflict of interest.