OpenAI’s latest push to make ChatGPT a “personal agent” now reaches directly into users’ wallets, promising richer financial help while raising fresh questions about how much intimate data people should share with AI.

Early development and launch

On May 14, OpenAI announced “a new personal finance experience in ChatGPT” for U.S. Pro subscribers, enabling them to securely link bank, card, and investment accounts and see where their money is going in a single dashboard.¹ The company framed the rollout as a limited preview to “learn from real-world use” before expanding more broadly.¹

The next day, OpenAI detailed that the feature relies on Plaid, the financial data network used by thousands of apps, to handle secure connections to more than 12,000 U.S. institutions.² Once linked, users can view portfolio performance, spending, subscriptions, and upcoming payments, and ask questions ranging from “Has anything changed in my spending?” to “Help me build a plan to buy a house in five years.”³

Tech outlets quickly underscored the stakes. The Verge noted that “OpenAI now wants ChatGPT to access your bank accounts,” calling it “a test” of how much users trust AI with sensitive financial data.⁴ TechCrunch similarly highlighted that OpenAI’s new tools “will let you connect bank accounts” so the model can analyze balances, transactions, and investments.³

Industry context and OpenAI’s pitch

Coverage placed the move in a broader surge of AI-powered money management. AI Magazine reported that the share of Americans using AI to manage finances jumped from 10% to 55% in recent years, but only 18% are comfortable letting AI make major financial decisions — a “trust gap” OpenAI is trying to navigate with its Finances experience.²

OpenAI positions the tool as informational, not a replacement for human advisers, and stresses that users can disconnect accounts and delete “financial memories.”² The company also points to GPT‑5.5’s improved reasoning and an internal benchmark built with more than 50 finance professionals to argue that the model can handle nuanced, context‑heavy money questions.⁵

Co‑founder Greg Brockman cast the launch as “a further step towards ChatGPT becoming your personal agent, operating on your behalf 24/7.”Greg Brockman (gdb)⁶ CEO Sam Altman amplified the announcement by retweeting ChatGPT’s description of a “preview for Pro users” who can securely connect accounts and see their “full financial picture” inside the app.Sam Altman (sama)⁷

Privacy fears and regulatory questions

Skeptical voices quickly emerged. The Next Web warned that letting ChatGPT see bank and loan accounts gives OpenAI access to “the most intimate data category left,” arguing that while the pitch is convenience, “the risk is everything else” — from privacy to future regulatory scrutiny.⁵

The Verge highlighted specific tensions: OpenAI promises “control over their data,” including the ability to disconnect accounts and limit whether information is used to “improve the model,” but it may take up to 30 days for synced data to be fully deleted.⁴ That delay, critics suggest, could become a flash point if something goes wrong.

Competing visions of AI and money

Supporters see a logical evolution: hundreds of millions already ask ChatGPT about budgeting and investing; tying those queries to real account data could improve financial literacy and help more people access sophisticated guidance.² Plaid’s CEO called it “an important step towards bringing sophisticated, personalised financial advice to more people.”²

Skeptics, however, worry that the same integration could blur lines between neutral assistance and commercial incentives, especially if future partnerships — such as with Intuit — enable journeys from advice to product applications inside ChatGPT.²³ As OpenAI tests its finance feature with a small U.S. Pro cohort, that tension between convenience and control is likely to shape how quickly — and how widely — AI moves deeper into people’s financial lives.