AI, Probability, and Kurt Gödel

Abstract

Probability is central to the functioning of AI. But LLMs tend to regularly hallucinate, or respond to interaction or other inquiries in improbable ways. Nonetheless, through proper training AI can be transformed into an increasingly closed system of options, thereby creating more certainty and improving the decisions made by this technology. Scaling is the process touted regularly to achieve this aim. Nonetheless, the creation of closed systems is in question following the work on logic by Kurt Gödel. Subsequent to Gödel’s incompleteness theorem, the creation of any closed is in jeopardy. The impact of this discovery for understanding probability is important, including the functioning of AI. The implications of Godel’s work on logic and AI, particularly the ability of AI to more closely mimic human decision-making is examined to make AI more socially relevant.