This thesis conducted a longitudinal analysis examining the evolution of public and media perceptions towards Large Language Models (LLMs) from December 2022 to October 2024. Utilizing more than 300,000 Reddit submissions and approximately 10,000 news articles, it applied sentiment analysis and topic modelling to explore the trends of sentiment and topics in the context of LLMs. Public discussions on Reddit exhibited gradual increases and fluctuations tied to the releases of new models and contained a wide range of sentiments, such as approval, gratitude, disappointment, and annoyance. The amount of public discussions on Reddit steadily increased over time, with noticeable spikes in activity following the release of major LLMs. The dominant sentiment also varied across topics, for instance, computer science-related discussions showed relatively higher proportions of positive attitudes toward LLMs compared to other themes such as health or education. In contrast, media coverage sharply increased after significant events and maintained a neutral emotion in general. Topic analyses highlighted common discussion areas between the public and media, such as healthcare, education, and career. However, the public emphasized technical details and daily-life applications, whereas media reports centred around corporate events, ethical concerns, and broader societal implications.