Bachelor / Master - Projektgruppe or Lab - LLMs and ethical dilemmas

As AI becomes more integrated into decision-making processes, understanding how these systems handle ethical dilemmas is increasingly important.

In this lab, you will explore how different LLMs, including both open-source and closed-source commercial models, respond to various scenarios. Analyze the variables that influence their behavior, such as language model architecture, the nature of the training data, or the phrasing of the prompts.

You can choose one of two approaches to investigate this topic:

Option 1: Focus on how these models respond to moral dilemmas.

Option 2: Focus on how these models behave in simple competitive or cooperative games.

In both cases, the goal of this lab is a deeper understanding of AI behavior in ethically charged situations, and the identification of variables that influence how LLMs evaluate and react to them.

Prior experience with LLMs is recommended, but not required.