Mind over matter. The philosophical arguments around AI, natural intelligence and memory

The Centre for Independent Studies Research Collection

Mind over matter. The philosophical arguments around AI, natural intelligence and memory.

In this intriguing research paper by Professor John Sweller, he deftly navigates the complex intricacies surrounding artificial intelligence (AI), natural intelligence, and memory. The entry in the research series of the Centre for Independent Studies (CIS), he challenges traditional perspectives as he critically examines both artificial and natural systems and what denotes them as intelligent. Professor Sweller employs metaphor and insights from evolutionary biology to provide an innovative understanding of intelligence’s foundation.

Professor Sweller underscored the essential role of knowledge, especially in educational contexts focusing on students' learning process. He posited that human cognition, an intelligent system itself, largely depends on our existing knowledge. Furthermore, he emphasized that the importance of knowledge in human cognition has been often downplayed both in educational research and AI development. This subtle observation explains AI's remarkable growth in recent times.

In a remarkable comparison, he explored the idea that like humans who struggle with limited intelligence due to insufficient knowledge, AI applications too have limitations in their utility without considerable accessible knowledge and information, the recent proliferation of which has been facilitated by the massive storage of data.

He delves deeper into the potential limitations of AI and the inefficiency of discovery learning for both humans and artificial intelligence systems. This becomes particularly relevant when false or implausible conclusions, known as AI hallucinations, are generated due to insufficient or low-quality data available to an algorithm.

Sweller stimulatively invites us to grapple with the concept of intelligence, its interplay with our knowledge and the foundations of intelligent systems. By reflecting on these foundations, we can better navigate the opportunities, challenges, and limitations of today’s AI and its potential implications for future educational systems and wider societal dimensions.