As Anthropic approaches its initial public offering at a record valuation, the company has launched a new feature in its AI assistant Claude named "Personal Finance". This step comes as a direct competitor to the "Finances" feature launched by OpenAI in ChatGPT last May. Although both tools aim to help users understand their spending patterns, deep technical and philosophical differences separate them, especially regarding the privacy of sensitive financial data.
Points of Intersection: How AI Views Your Bank Account
Both companies share the underlying infrastructure for connecting accounts; Anthropic and OpenAI both rely on the well-known Plaid platform to secure connections with major financial institutions such as Bank of America and Capital One. These features operate on a "read-only" permission, meaning the AI can view and analyze data but does not have the ability to execute transfers or modifications on accounts.
Functionally, both bots can answer general financial questions, such as planning for a home purchase or budgeting for a family vacation, while providing visual dashboards that illustrate monthly expenses, investments, and spending habits. Anthropic specifies the data "Claude" sees as the bank name, account type, balance, credit limits, and transaction history for up to two years—data similar to what OpenAI requires to run its feature.
Privacy: The Major Gap Between Anthropic and OpenAI
Privacy is the primary differentiator through which Anthropic attempts to attract users who are hesitant to share their financial data. By default, Anthropic has announced that conversations taking place within the personal finance section in "Claude" are not used to train its AI models. The company also confirmed that it stops seeing any previously provided data as soon as the user disconnects their bank account.
In contrast, OpenAI follows a different approach; financial conversations are subject to the same general training settings in ChatGPT. This means your financial data may be used to improve the model unless the user manually goes into data control settings and disables the "Improve the model for everyone" option. If a user decides to disconnect their accounts, OpenAI retains the synchronized data in its systems for up to 30 days before permanently deleting it.
Business Models and Future Directions
Differences also appear in how these services are monetized. Anthropic has so far maintained an ad-free and sponsored-recommendation-free model, confirming that it does not offer its own loans or credit cards. OpenAI, on the other hand, has already begun testing ads in its lower-cost subscription tiers and plans to expand its financial features through a partnership with Intuit to include services like credit card applications and tax advisory.
What Practically Changes for the User?
This development represents a shift in the role of the AI assistant from a mere text conversationalist to a financial manager privy to the finest details of personal wealth. For the user, the choice between the two platforms depends not only on the quality of the analysis, but on the level of trust in how this data is handled:
- For Claude users: The focus is on strict privacy and the non-use of data in training, making it a preferred choice for professionals and those seeking a neutral analytical tool.
- For ChatGPT users: The advantage lies in future integration with broad tax and credit services, but it requires greater awareness of privacy settings to ensure data does not leak into model training processes.
Ultimately, these tools remain in an initial advisory stage, with both companies emphasizing that they do not replace a human financial advisor, but rather act as a smart mirror reflecting the user's financial reality and helping them make more informed decisions based on their historical data.