Sentiment is an important concept in economics and finance and has been the focus of many studies. Individual investors, professional investors, corporate managers, and households have sentiments on the economy and financial markets which affect their decisions, and hence economic activities and asset prices. Measuring sentiment and determining what factors affect it have significant importance in finance research. My dissertation studies this subject by introducing state-of-the-art methods from artificial intelligence to measure the sentiment in several sources of business text data, that is, public firms’ disclosures and mutual funds’ reports. I investigate the information content, determinants, and the effects of the sentiments on asset prices and investment decisions of investors. In chapter one, we use a novel text classification approach from deep learning to accurately measure sentiment in a large sample of 10-Ks. In contrast to prior literature, we find that both positive and negative sentiments predict abnormal returns and abnormal trading volume around the 10-K filing date and future firm fundamentals and policies. Our results suggest that the qualitative information contained in corporate annual reports is richer than previously found. In chapter two, I study the sentiment of mutual fund managers towards the stock market. Using a direct measure of managers’ market expectations extracted from mutual funds’ semi-annual reports, I find that fund managers extrapolate their funds’ past performance into their market outlook. Funds with managers who have higher market expectation take more risk by increasing their equity holdings and the beta of their equity portfolios, but underperform subsequently. In chapter three, we study the sentiment of mutual fund managers about specific stocks in their portfolios. We study some mutual funds’ practice of voluntarily disclosing investment ideas in their annual reports. The practice involves, at a minimum, expressing views on stocks which fund managers are optimistic about. We find that managers of larger and better performing funds discuss positions that have recently underperformed, those that make up larger portions of their portfolios, and those they have held for longer periods. Our findings suggest that managers disclose these recommendations to boost their own fund performance and to attract additional capital.