The field of Large Language Models (LLMs) is rapidly advancing, with a growing focus on ethics and fairness. Recent research has highlighted the need for more reliable evaluation metrics and methods to detect subtle biases in LLM outputs. Several studies have proposed novel frameworks and datasets to assess the moral reasoning capabilities of LLMs, including the ability to recognize and explain hate speech, toxic content, and morally ambiguous scenarios. Additionally, researchers have explored the use of LLMs in real-world applications, such as organizational research and engineering education, and have identified potential gaps in current LLMs' ability to internalize and reflect human moral reasoning. Notably, some papers have shown that LLMs can exhibit human-like motivated reasoning and violent tendencies, which can be influenced by demographic factors and persona assignments. Overall, the field is moving towards a more nuanced understanding of the limitations and potential risks of LLMs, and the development of more robust and fair evaluation methods. Noteworthy papers include: PRISON, which proposed a unified framework to quantify LLMs' criminal potential across five dimensions. MFTCXplain, a multilingual benchmark dataset for evaluating the moral reasoning of LLMs via hate speech multi-hop explanation. Quantifying Fairness in LLMs Beyond Tokens, which introduced a novel statistical framework to evaluate group-level fairness in LLMs by detecting subtle semantic differences in long-form responses.