Realtime AI News
Citi tells customers using third-party AI agents: you're on your own, even if you get hacked
Citibank changed its online user agreement this week to make customers responsible for anything that goes wrong when they give third-party AI tools access to their accounts, including losses after a credential leak or hack. JPMorgan and Bank of America adopted similar policies last month, PNC followed the same day, and Citizens emailed customers a comparable warning.

Citibank, the third-largest US commercial bank, changed its online user agreement this week to tell customers that if they hand account access to third-party AI agents, they bear responsibility for whatever happens, even if the account is hacked.
A Citi spokesperson framed the update as disclosure. As third-party AI tools become more widely available, the bank said it wants customers to understand the risks of sharing banking credentials or authorizing those tools to access their accounts, and the revised agreement reinforces that customers who grant such access are responsible for activities conducted through those agents.
Citi's language goes beyond assigning blame for the agent's own transactions: if the AI tool leaks credentials or is compromised, the customer is on the hook for what attackers do with the account. That extends the liability shift from bad automation to outright security breaches.
The move is part of a pattern rather than a one-off. JPMorgan and Bank of America, the two largest US banks, announced similar policies last month; PNC, the eighth largest, changed its terms the same day as Citi; and Citizens emailed customers this week with a comparable warning, according to a message seen by TechCrunch.
The products in question range from frontier lab models to purpose-built personal finance tools. Such agents are typically given credentials to bank accounts and other financial portals so they can analyse spending and investment decisions, offer advice, and in some cases execute transactions and pay bills.
AI labs are already adjusting to the chill. Anthropic's personal finance tool, launched earlier this month, is explicitly billed as read-only and cannot move, manage or invest money, a design choice that doubles as a response to how institutions are reacting.
Banks have two motives. One is cybersecurity: unreliable automated actors inside account systems create risks that are hard to price. The other is competition, since agents that help consumers dodge unnecessary fees or low-interest deposits threaten the margins that banks earn on deposits, a prospect that briefly rattled markets earlier. Citi notes that its own AI agents can be used without additional liability.
What to watch is whether more institutions copy the language, and whether shared standards for agent identity and permissioning emerge. For now, the liability line is being drawn in bank terms of service rather than in a common regulatory framework.
Why it matters
By externalizing agent risk through contracts, banks lower their own legal exposure in the short term, but they also slow consumer adoption of personal finance agents and raise the trust bar to the point where only read-only products can clear it.
Nearby Updates
All11/19, 01:47
Google launches Stitch Loop, an autonomous product agent that keeps working after launch
Google announced Stitch Loop on November 18, a Google Labs experiment that turns Stitch into a long-running autonomous product agent connecting GitHub and live product signals. Built on Gemini 4 Carbon and CodeMender, it proposes fixes, builds them in a sandbox and opens pull requests for human review, with a global waitlist now open and priority for Google AI Ultra subscribers.
11/19, 01:54
Dubai approves 'City AI Agent' initiative to rebuild government operations around AI
Sheikh Hamdan bin Mohammed approved Dubai's 'City AI Agent' initiative at the AI Retreat during Dubai Future Week 2026 on November 18, aiming to make AI agents integral to how government anticipates needs and coordinates services. He also launched an ecosystem of 93 government-resource agents expected to handle over 34 million tasks and save more than 812,000 working hours a year across 85 entities.
11/19, 01:05
Team Turing launches Pie, a proactive AI agent for college students
Team Turing has launched Pie, an AI agent aimed at college students that is designed to surface tasks before the user asks. The product's selling point is proactivity rather than conversation, with campus life as its first target market.
11/19, 02:01
Sustainalytics flags high ESG risk at Anthropic, putting AI governance in focus
A Sustainalytics assessment carried by Morningstar places Anthropic in a high ESG risk category, tying the rating to questions about AI governance. The move shows how conventional investment analysis is starting to evaluate model developers on how they govern their own technology.