Reasoning about what to believe is often analyzed using probabilities, while reasoning about what to do combines probabilities with utilities.  A systematic alternative is provided by coherence theories of belief and decision, in ways that are more psychologically plausible and computationally efficient than probabilities. Nevertheless, powerful new AI models seem to use neither probability or coherence. What does their success tell us about rational thinking?

References

Thagard, P. (2024). Can ChatGPT Make Explanatory Inferences? Benchmarks for Abductive Reasoning. arXiv. Retrieved from https://arxiv.org/abs/2404.18982

Thagard, P. (2012). The cognitive science of science: Explanation, discovery, and conceptual change. Cambridge, MA: MIT Press. https://direct.mit.edu/books/monograph/3753/The-Cognitive-Science-of-ScienceExplanation

Thagard, P. (2024). Falsehoods fly: Why misinformation spreads and how to stop it. New York: Columbia University Press. https://cup.columbia.edu/book/falsehoods-fly/9780231213943/