As we can see the social media platforms such as Facebook, Twitter and Instagram have been growing in a vast and rapid form which have created an enormous volume of textual data that shows and reflects public emotions, opinions and behavioural patterns. To understand users' sentiment at scale by analyzing this data offers valuable insights for organizations, governments and researchers. Sentiment analysis with Artificial Intelligence (AI) is a crucial technique used for interpreting the emotional tone embedded within online conversations. In this paper a comprehensive examination of AI-driven sentiment analysis techniques is used to evaluate user emotions on social media. It shows the capabilities of machine learning and deep learning models, including Naive Bayes, Support Vector Machines (SVM), LSTM networks and transformer-based architectures such as BERT and GPT. This paper also explores the ongoing challenges faced when analyzing informal, diverse and contextrich social media data. Issues including sarcasm, irony, multilingual content, slang, abbreviations and rapidly evolving trends make sentiment classification hard to understand and require more robust AI models. Apart from technical aspects, this paper shows the wide-ranging applications of sentiment analysis in domains like marketing, politics, public health, brand monitoring and customer experience management. These applications demonstrate how sentiment insights support decision-making, enhance user engagement and help organizations react to public sentiment in real time. Furthermore, the paper discusses emerging research directions, including multimodal sentiment analysis that incorporates images and audio, fine-grained emotion recognition, cross-lingual models and explainable AI approaches that improve transparency and trust in AI-driven sentiment systems. Overall, this research highlights how AI continues to transform sentiment analysis on social media, offering new opportunities for understanding user behaviour and emotional expression in the digital age.