Major US Banks Disclaim Liability for AI Agent Transactions, Shifting Risk to Customers
Citibank, JPMorgan, Bank of America, and others updated terms of service to hold customers responsible for losses caused by third-party AI tools accessing their accounts.
Last verified:
Five major US financial institutions have updated their customer agreements to explicitly disclaim responsibility for losses caused by third-party AI agents granted access to bank accounts. The policy shift places full liability on the customer, even if the AI tool itself is breached or its credentials are exposed.
Citibank’s Terms Update Reflects Broader Bank Resistance
According to TechCrunch AI, Citibank, the third-largest commercial bank in the United States, modified its Online User Agreement this week to inform customers authorizing third-party AI tools to access their accounts that they accept all associated risks. A Citi spokesperson explained to TechCrench that the bank wanted to “provide greater clarity about those risks and to reinforce that customers who choose to grant account access to third-party AI agents are responsible for activities conducted through those agents.”
TechCrunch reports that JPMorgan Chase and Bank of America, the nation’s two largest banks, announced comparable liability disclaimers in October. PNC Bank, the eighth-largest, updated its terms on the same day as Citibank. Citizens Bank similarly notified customers this week via email that they bear responsibility for AI agent actions on their accounts.
Scope of Customer Liability
The liability disclaimers extend beyond authorized transactions executed by malfunctioning AI tools. TechCrunch notes that Citibank’s updated terms hold customers responsible if a third-party AI agent leaks banking credentials or is itself compromised, with any subsequent fraudulent activity treated as the customer’s loss to bear. This creates a scenario where a customer who grants an AI agent legitimate access could face financial consequences from a breach of the agent’s infrastructure—not the bank’s systems.
The Competitive Calculation Behind Liability Shifts
Banks have both security and market-share motivations for resisting third-party AI agents in their systems. Financial institutions are concerned that autonomous agents could help customers identify and avoid fees or redirect deposits to higher-yield accounts, eroding profit margins historically dependent on consumer inattention. By disclaiming liability, banks reduce their exposure while making third-party AI access less attractive—and less safe—for consumers.
Anthropic’s approach to this landscape illustrates one alternative. TechCrunch reports that the company’s personal finance tool, launched earlier in November, is explicitly designed as read-only, unable to move, manage, or invest customer funds, thus eliminating the account-access liability problem entirely.
Why This Matters
As AI startups pursue growth in fintech, the asymmetric liability structure creates a friction point for adoption. Customers who want AI assistance with financial decisions now face a binary choice: use a bank’s proprietary AI tool (if available) with standard liability protections, or accept full financial and security risk by deploying a third-party agent. This may slow the uptake of AI agents in personal finance and reinforce banking incumbents’ control over the customer interface, even as it formally disclaims their responsibility for third-party breaches. Regulators monitoring AI-driven financial services may eventually scrutinize whether this liability shift complies with consumer protection standards, particularly in cases where the AI tool’s vulnerability, not customer negligence, enables fraud.
Frequently Asked Questions
What happens if an AI agent I authorized gets hacked and someone steals money from my account?
According to the new terms of service, you are responsible for the loss. Banks including Citibank and JPMorgan Chase have updated their policies to disclaim liability for unauthorized activity conducted through third-party AI agents, even if the agent itself was compromised.
Do banks' own AI tools carry the same liability risk?
No. Citibank's terms note that AI agents the bank provides can be used without additional liability beyond standard account protections. The liability shift applies only to third-party AI tools.
Why are banks making this change now?
As AI startups develop tools that can access bank accounts to analyze spending and handle transactions, financial institutions view both a security risk and a competitive threat—AI agents could help customers avoid fees or identify better investment options.
Is there a safer way to use AI for personal finance?
TechCrunch notes that Anthropic's personal finance tool, launched in November 2026, is designed as read-only and cannot move, manage, or invest funds, eliminating the account-access risk.