Human Brains Outperform AI in Skill Transfer, New Research Reveals
Table of Contents
A new study published in Nature demonstrates the remarkable flexibility of the human brain – and its superiority to current artificial intelligence – in applying learned skills to new challenges. Researchers have identified how the brain repurposes existing neural structures,likened to “cognitive legos,” to rapidly adapt and learn,a capability that remains elusive for even the most advanced AI models.
The ‘Cognitive Lego’ breakthrough
The research, led by a team at Princeton University, didn’t directly study human brains.Instead, scientists focused on rhesus macaques (Macaca mulatta), whose brains share important biological and functional similarities with our own. These monkeys were tasked with identifying shapes and colors on a screen, indicating their responses through eye movements. Throughout the experiments, brain scans meticulously tracked neural activity, revealing a surprising pattern.
The scans showed the monkeys utilizing distinct “blocks” of neurons – what the researchers playfully termed ‘cognitive Legos’ – across different tasks. These pre-existing neural blocks weren’t created anew for each challenge; rather, they were efficiently repurposed and recombined. This demonstrates a level of neural flexibility that currently surpasses the capabilities of artificial intelligence.
“State-of-the-art AI models can reach human, or even super-human, performance on individual tasks,” explains neuroscientist Tim Buschman, from Princeton University. “But they struggle to learn and perform many different tasks.”
How the Brain Builds New Skills
The study found that the brain’s flexibility stems from its ability to reuse fundamental components of cognition. By “snapping together” these ‘cognitive Legos,’ the brain constructs solutions to novel problems. This process allows for rapid adaptation and learning, leveraging existing knowledge instead of starting from scratch.
Researchers observed that these cognitive building blocks were largely concentrated in the brain’s prefrontal cortex, a region crucial for higher-level cognitive functions like problem-solving, planning, and decision-making.furthermore,activity within these blocks diminished when they weren’t required,suggesting the brain efficiently “files away” unused neural components to focus on the task at hand.
“I think about a cognitive block like a function in a computer program,” Buschman elaborates. “One set of neurons might discriminate color, and its output can be mapped onto another function that drives an action. That association allows the brain to perform a task by sequentially performing each component of that task.”
Implications for AI and Neurological disorders
The findings have significant implications beyond basic neuroscience. researchers believe this understanding of brain function could inform the growth of more adaptable AI systems. Current AI models often suffer from catastrophic forgetting – the inability to learn new tasks without losing proficiency in previously learned ones. Mimicking the brain’s ‘cognitive Lego’ approach could overcome this limitation.
The research also holds promise for treating neurological and psychiatric disorders where individuals struggle to apply learned skills to new situations. Understanding how the brain builds and reuses cognitive components could lead to targeted therapies to enhance cognitive flexibility.
The animals in the study were required to discriminate between shapes and colors in three separate but related tasks, continually applying previously learned information – a key aspect of the research.
ultimately, these “cognitive Legos” highlight a fundamental difference between the human brain and current AI models: our brains are inherently adaptable, capable of leveraging past experiences to navigate new challenges. This adaptability is a key element of intelligence that artificial intelligence has yet to fully replicate.
“If, as suggested by our results, the brain can reuse representations and computations across tasks, than this could allow one to rapidly adapt to changes in the habitat, either by learning the appropriate task representation through reward feedback or by recalling it from long-term memory,” the researchers conclude.
