Decision Making with Gen AI and LLMs. Case Study: Automatic decision-making for bots and machines

This paper explains decision-making using Soft Sets and Soft Rough Sets, keeping in view the changing landscape of NLP and LLMs in the arena of Machine Learning. Here, the queries are natural-language queries, represented as vectors, with and without a Gen AI, to compute the Soft Set. Gen AI can be used to generate a natural-language description of an object in the Universe. Apart from this, this paper also discusses, with examples, decision-making with such concepts and their application to bot decision-making. Note that decision-making for bots must be carefully developed, as we want machines to explain their decisions (explainable AI) to prevent wrong acts. This paper demonstrates that a plethora of decision-making algorithms can be applied to machine applications, given the many algorithms in our research books. It’s time to take these decision-making algorithms out of research papers and put them to work with bots, with due care and with final authority resting with the human owner of the machines.

(PDF) Decision Making with Gen AI and LLMs. Case Study: Automatic decision-making for bots and machines. Available from:

https://www.researchgate.net/publication/396386224_Decision_Making_with_Gen_AI_and_LLMs_Case_Study_Automatic_decision-making_for_bots_and_machines [accessed Oct 10 2025].

This paper presents a way for bots to understand the decisions humans want from them. Many decision-making algorithms have been studied in decision science and management sciences. It’s time to move thesedecision-making algorithms from books into practical use. It’s time to make decision-making algorithmswork with Gen AI (LLM and more). We want bots to do tasks we can’t, such as cleaning a cricket stadium ona hot summer day. This can be done by bots first finding and studying the cricket ground, then labeling eacharea using their computer vision skills. These skills can convert images into natural language, which can bevectorized or fed back into the system to identify clusters of objects (areas that need cleaning). Once weobtain clusters of objects, we generate a Soft Set, which can be used in decision-making algorithms such asSoft Rough Sets. It can also be used to compute the spanning set of the problems. However, it must beensured that machines using decision-making algorithms do so with the transparency of the human ownerand that the decisions are safe and not harmful. This is one side-effect of autonomous machine decisions that machines can create the aim to destroy mankind if we do not look at the decisions that machines need totake. It must be noted that the field of decision science leverages the fact that machines need to automate this decision-making task, while humans must sign off on it.