The venture capital firm wants more institutional investment in European technology companies and backs efforts to simplify cross-border business.
Citibank puts account risks from third-party AI agents on customers
The bank’s updated terms extend customer responsibility to activity resulting from compromised AI tools. Other major banks have adopted similar policies.

Citibank has revised its online banking agreement to make customers responsible for account activity carried out through third-party artificial intelligence agents, including activity by hackers if those tools expose login credentials or are compromised.
A Citi spokesperson said the changes clarify the risks of giving outside AI tools banking credentials or account access. The bank’s terms distinguish between outside tools and agents supplied by Citi, which customers can use without taking on additional liability.
JPMorgan and Bank of America have announced comparable policies, while PNC has also revised its terms. Citizens warned customers by email that they would be responsible for actions taken by their AI assistants.
AI agents can use access to bank accounts and other financial services to examine spending, provide financial advice or carry out payments and transactions. Banks adopting these policies say they will not take responsibility for reversing transactions made by automated assistants, whether intentional or accidental.
Some AI providers are limiting what their financial tools can do. Anthropic’s recently launched personal finance tool allows users to read financial information but cannot transfer, manage or invest their funds, according to the company.
