Imagine having an AI assistant that could condense complex legal judgments into concise summaries, saving lawyers countless hours of tedious work. That's the promise of recent research exploring the application of large language models (LLMs) and other AI techniques to legal case summarization. Traditionally, extracting key sentences has been the go-to method, but now, researchers are investigating more sophisticated approaches. This research delves into the potential of various models, including cutting-edge LLMs like ChatGPT and specialized legal AI models. The results, tested on UK and Indian Supreme Court cases, reveal that AI can indeed generate more natural and coherent summaries than traditional methods. However, there's a catch: these AI-generated summaries sometimes contain factual inconsistencies or even fabricate information—a phenomenon known as 'hallucination.' For example, an AI might get a monetary value wrong or invent a non-existent legal statute. Researchers are actively working on solutions, such as fine-tuning AI models on specific legal domains and crafting smarter prompts to guide the AI. One promising strategy involves using semantic similarity: when the AI hallucinates a name or number, the system finds the most similar entity in the original document and replaces the error. While there are still challenges to overcome, this research suggests that AI-powered legal summarization could soon become a valuable tool, though likely with a human expert in the loop to ensure accuracy.