Citibank has updated its terms of service to explicitly assign responsibility to customers for actions taken by third-party AI agents, even in cases involving compromised credentials. According to Decrypt, the financial institution is shifting the burden of liability for “rogue” artificial intelligence interactions directly onto users. This policy update addresses scenarios where external AI tools interact with banking systems, potentially causing unauthorized transactions or account modifications through leaked login information.
The revised terms clarify that customers are liable for any consequences resulting from their AI agents’ operations. This includes situations where hackers exploit leaked credentials to manipulate AI assistants into executing fraudulent activities. By establishing this framework, Citi aims to mitigate risks associated with the growing integration of autonomous AI tools in financial services. The bank argues that users must maintain strict control over their AI interactions to prevent misuse.
Implications for consumers are significant, as they now bear the financial risk for AI-driven errors or hacks. This stance contrasts with traditional banking models where institutions typically absorb losses from unauthorized access due to security failures. The update reflects a broader trend in fintech, where liability boundaries between users, platforms, and AI developers remain unclear. Legal experts suggest this move may challenge existing consumer protection laws. Customers are advised to review the new terms carefully, as acceptance is mandatory for continued service. The decision highlights the urgent need for regulatory clarity regarding AI accountability in banking. Without clear guidelines, users face potential financial exposure from complex AI interactions. Citi’s approach sets a precedent that could influence other major banks navigating similar technological challenges.
