Policy

Bank of America warns customers against sharing banking credentials with AI assistants

BofA urges customers not to hand over login credentials to AI agents, citing unauthorized account access risks and customer liability for fraudulent activity.

Last verified:

Bank of America’s credential-sharing warning

Bank of America issued a customer advisory titled “Protecting your finances in the AI World,” explicitly urging customers never to share banking usernames and passwords with AI-enabled tools or agents that claim to act on their behalf, according to TechCrunch AI. A Bank of America spokesperson confirmed the email’s authenticity after it circulated on social media.

The timing reflects a growing disconnect between consumer adoption of AI agents and financial-sector readiness. Customers are increasingly handing over login credentials to personal AI assistants as the tools proliferate, even as banks remain uncertain how to safely integrate agent-based access into their platforms.

The risks of credential sharing

Three distinct threats emerge from sharing banking credentials with AI assistants. First, an agent could execute transactions without explicit per-transaction authorization—customers might discover unauthorized purchases or transfers only after the fact. Second, fraud-detection systems at banks often lock accounts when they detect anomalous activity, including access from unfamiliar devices or IP addresses where an AI agent might be running, potentially locking legitimate account owners out of their own funds. Third, and most severe, a security vulnerability in the AI assistant itself could expose banking credentials to financially motivated attackers.

According to TechCrunch AI, Bank of America’s email emphasized that customers remain responsible for any account activity resulting from credentials they voluntarily share. This liability structure means that if an AI agent makes unauthorized purchases or a security flaw leaks credentials to hackers, the customer—not the bank—bears the financial loss.

The broader banking challenge

Banks and payment processors are still negotiating how AI agents should interact with customer accounts. According to TechCrunch AI, industry players are reportedly struggling with unauthorized agents accessing customer money while simultaneously exploring legitimate use cases like autonomous payment agents that buy goods at optimal prices or times. The tension illustrates a gap between what consumers want (AI assistants that can transact autonomously) and what financial institutions can safely allow without opening themselves to liability or regulatory scrutiny.

Why This Matters

Bank of America’s warning signals a structural misalignment in the emerging AI-agent economy. As long as customers remain legally liable for credentials they share and AI assistants lack formal authorization mechanisms, most banks will discourage or prohibit credential-sharing rather than build legitimate integrations. This creates an incentive for customers to either avoid using AI agents for financial tasks or to use unauthorized workarounds—precisely the scenario banks most want to prevent. Financial services firms that want to capture the agentic-payment opportunity will need to develop authenticated delegation models (similar to OAuth for APIs) rather than rely on password-sharing. Until then, expect more warnings from BofA and its peers.

Frequently Asked Questions

Why is Bank of America warning customers now?

Customers are increasingly sharing banking credentials with AI assistants as personal AI agents become more popular. Banks are grappling with unauthorized agents accessing customer accounts, and account security vulnerabilities could expose credentials to hackers.

Who is liable if an AI assistant misuses my banking credentials?

According to Bank of America's terms, customers are responsible for any activity resulting from credentials they voluntarily share. This means if an AI agent makes unauthorized transactions or a security flaw leaks your password, you bear the financial responsibility.

What could go wrong if I share my banking password with an AI assistant?

Risks include unauthorized transactions, account lockouts from fraud-detection systems, security breaches that expose credentials to hackers, and potential financial loss with limited recourse under most bank account agreements.

#ai-agents #banking #security #customer-liability #financial-services