LONDON: Rapid advances in artificial intelligence (AI) are challenging long-held assumptions about human intelligence, but experts say human minds will continue to remain distinct rather than be replaced.
Writing in a recent analysis, cognitive scientist Tom Griffiths argues that concerns about AI surpassing human intelligence are based on a flawed comparison. “Talking about superhuman AI assumes that intelligence is a single scale,” he said, adding that this perspective overlooks the diversity of ways intelligence can be expressed.
Griffiths compares the debate to measuring height, noting that while people may fear being overtaken, intelligence is fundamentally different. Unlike height, which follows a single dimension, intelligence varies across contexts, environments and biological constraints.
He explains that human cognition has evolved under limitations such as short lifespans, finite brain capacity and restricted communication abilities. These constraints, rather than being weaknesses, have shaped unique human strengths, including learning from limited experience, creativity and social cooperation.
By contrast, AI systems operate under entirely different conditions. They can process vast amounts of data, expand computing capacity and share information instantly across networks. However, Griffiths points out that these advantages do not make AI universally superior.
“Human intelligence is a response to our limitations,” he writes, highlighting that machines often rely on massive datasets and computational power, whereas humans can generate insights with far less information. He also notes that AI systems sometimes struggle with tasks that humans find straightforward, such as interpreting numbers or context accurately.
The analysis suggests that humans and AI will continue to develop different forms of intelligence shaped by their respective environments and constraints. Rather than replacing human abilities entirely, AI is expected to complement them in specific areas.
Griffiths concludes that intelligence should not be viewed as a competition but as a spectrum of capabilities, with humans and machines excelling in different domains.
