OpenAI is expanding the capabilities of ChatGPT to include direct access to users' bank accounts through a new feature in preview. Announced today, this integration allows users to securely link their financial accounts with Plaid, a platform used by over 12,000 institutions such as Chase, Fidelity, and Capital One. According to OpenAI, more than 200 million monthly users already ask for financial advice, and this update aims to provide a comprehensive view of their financial standing tailored to personal goals and lifestyle priorities. Once connected, the chatbot enables users to view spending history dashboards, track active subscriptions, and receive guidance on major decisions like purchasing a home or applying for credit cards. It can also flag unusual changes in spending habits. However, the service is not designed to execute transactions. OpenAI explicitly stated that ChatGPT cannot make changes to bank accounts or access full account numbers, though it can view balances, transactions, stock portfolios, and liabilities such as mortgages and credit card debt. The feature will initially roll out only to users in the United States subscribed to the ChatGPT Pro plan, which costs $200 per month. Following initial feedback, the company plans to expand availability to Plus subscribers and eventually to all users. This move follows the January launch of ChatGPT Health, a similar service for health-related queries that emphasizes the need for user trust regarding data privacy. To address security and privacy concerns, OpenAI promises users full control over their data. Users can disconnect their bank accounts at any time and manage saved financial memories, such as goals or obligations. The company allows users to opt out of having their conversation data used for AI training. However, data retention policies allow up to 30 days for full deletion from OpenAI systems after a request. Critics and analysts note that while the feature offers convenience, the company does not yet specify how it will monetize this vast financial data beyond model training or detail specific protections against potential system hacks. The rollout represents a significant test of user confidence in AI handling sensitive financial information.