The rise of digital media has led to the rapid spread of financial misinformation, posing significant risks to investors, markets, and economic stability. As financial markets heavily rely on accurate information for decision-making, misinformation can lead to severe consequences, including stock price manipulation, investor losses, and regulatory failures. Traditional fact-checking methods are often slow and inefficient in handling the large-scale and rapidly evolving nature of financial misinformation. Consequently, machine learning and artificial intelligence (AI) have emerged as powerful tools to detect and mitigate financial misinformation effectively. This paper explores recent advancements in financial misinformation detection using large language models (LLMs) (Sultana et al. in J AI Financ 29:332–350, 2024) and deep learning-based classification models. We review different methodologies, including instruction-tuned LLMs, misinformation measurement frameworks, and machine learning-based classification models. Specifically, we analyze the effectiveness of models such as FMDLlama, which leverages instruction-tuning for detecting and explaining financial misinformation, and structured misinformation measurement frameworks that quantify misinformation at the firm level. Additionally, we discuss the challenges associated with financial misinformation detection, such as the lack of standardized datasets, the evolving nature of misinformation, and the interpretability of AI models. By comparing existing methodologies and presenting insights into their strengths and limitations, this paper highlights the importance of structured datasets, AI-driven analysis, and robust evaluation benchmarks in improving financial misinformation detection. Our findings provide a foundation for future research to develop more accurate, scalable, and interpretable misinformation detection systems that can enhance financial market integrity and investor confidence.