Understanding Human Cognition in AI Systems

The logos, ethos, and pathos of your LLMs

Tom discusses the challenges of understanding human cognition in AI systems, highlighting the implications of probabilistic AI thinking. The Laws of Thought details the quest to use mathematics to describe human thought, from its origins to modern AI systems.

Why This Matters

The technical reality of AI systems is that they differ significantly from human minds, despite being designed to mimic certain aspects of human cognition. This disparity can lead to failures in AI decision-making, emphasizing the need for a deeper understanding of human thought processes and their mathematical descriptions. The cost of such failures can be substantial, as seen in instances where AI systems have made suboptimal decisions due to their limited understanding of human context and ethics.

Key Insights

• The Laws of Thought aims to mathematically describe human cognition, originating over three hundred years ago and influencing modern AI systems.
• Aristotle’s philosophical discussions on consciousness and sentience are relevant to current AI debates, as highlighted by Tom’s work.
• Neural networks in AI differ from human minds, as discussed in the context of The Laws of Thought and its implications for AI development.

Practical Applications

• Use case: Implementing AI systems that mimic human decision-making processes, such as those used in expert systems. Pitfall: Overreliance on probabilistic thinking without considering ethical implications, leading to potential biases in decision-making.
• Use case: Developing AI-powered chatbots that understand human emotions and respond appropriately. Pitfall: Failing to account for the complexity of human emotions, resulting in inappropriate or insensitive responses.

References:

• https://stackoverflow.blog/2026/02/10/the-logos-ethos-and-pathos-of-your-llms/