AI apps, like ChatGPT, Grammarly, and Google Bard, are rapidly becoming more popular in the enterprise. However, their use comes with the risk of accidentally sharing sensitive information. The most frequently shared sensitive data on ChatGPT, the leading AI app, is source code. On average, an organization sees 183 incidents of source code being posted to ChatGPT per 10,000 enterprise users per month. This report examines how organizations are balancing the benefits of AI tools while also managing the associated risks, highlighting an increasingly popular strategy that involves DLP and interactive user coaching.