Mind the AI Gap: Understanding vs. Knowing

– First Published: May 2024 –

Gorkem Turgut (G.T.) Ozer

Paul College of Business and Economics

March, 2026

Before we start…

  • I approach this not as a philosopher or cognitive scientist, but as a professor of Information Systems, and focus on how LLMs can augment human learning when the goal is to understand vs. know.

  • First, we need a definition. In this talk, knowledge is information-produced belief. One knows that p if and only if one’s belief that p is caused (or causally sustained) by the information that p.1,2

  • Hence, knowing doesn’t require reasoning. The information flow does all the work. Understanding, however, requires reasoning.

    • One can do things without understanding: e.g., drive a car without understanding the mechanics of motion. One can also use tools to perform tasks: e.g., a student can use a calculator to solve equations without grasping the underlying algebra.

Opening questions

  1. Can LLMs help us identify and fill our knowledge gaps?
  2. Do LLMs think and understand?
    • What does it mean to understand?
    • What does it take to understand?
  3. How can LLMs help us understand topics and acquire new skills, and move beyond mere knowledge and execution?
  4. In closing: Is AI still just a tool complementing human mind? Or is it a mind now?

Strong AI vs. Weak AI3

Strong AI is associated with the claim that an appropriately programmed computer could be a mind and could think at least as well as humans do.

  • Machines that matches or surpasses human intelligence

Weak AI is associated with attempts to build programs that aid (complement/augment), rather than duplicate (substitute/automate), human mental activities.

  • Machines that handle tasks once performed by humans

Deductive reasoning with GPT-4o (or lack thereof)10

* GPT fails in a basic deductive reasoning task. Click on the images to zoom in.

Deductive reasoning with GPT-o1 (or lack thereof)11

* GPT fails in a basic deductive reasoning task. Click on the images to zoom in.

IN CLOSING:
Is today’s AI a tool or a mind?

Footnotes

  1. Dretske, F. I. (1981). Knowledge and the Flow of Information. MIT Press. p. 86

  2. Jäger, C. (2004). Skepticism, information, and closure: Dretske’s theory of knowledge. Erkenntnis, 61(2), 187-201.

  3. Nilsson, N. J. (2009). The quest for artificial intelligence. Cambridge University Press. Originally introduced in Searle, J. R. (1980). Minds, brains, and programs. Behavioral and brain sciences, 3(3), 417-424.

  4. Searle, J. R. (1980). Minds, brains, and programs. Behavioral and Brain Sciences, 3(3), 417-424.

  5. Anderson, D. L., Stufflebeam, R., & Cox, K. (2018). Searle’s Chinese Room Argument. Illinois State University.

  6. Christian, James Lee. (1990). Philosophy: An Introduction to the Art of Wondering. Holt, Rinehart & Winston.

  7. Feynman, R. P. (1988). What Do You Care What Other People Think? W. W. Norton & Company.

  8. Kant, I. (1908). Critique of pure reason. 1781. Modern Classical Philosophers, Cambridge, MA: Houghton Mifflin, 370-456.

  9. Van Hoeck, N., Watson, P. D., & Barbey, A. K. (2015). Cognitive neuroscience of human counterfactual reasoning. Frontiers in Human Neuroscience, 9, 420.

  10. Queries were executed using gpt-4o-2024-05-13 on May 22, 2024.

  11. Queries were executed using gpt-o1-preview on September 13, 2024.

  12. This does not imply unbiased or complete knowledge. The issues of bias and ethics are outside the scope of this discussion.