# U.S. Banks Block Meta's Muse AI from Executing Financial Transactions Major financial institutions restrict AI agents to prevent unauthorized account access and liability risks. - **Source:** [Gizmodo](https://brifing.si/source/63382431-cef0-4c41-883e-2f88dc21351e) - **Published:** 2026-10-16 - **Topic:** [Finance](https://brifing.si/c/finance) - **Original article:** - **On Brifing:** https://brifing.si/s/d064a5ec-e07e-4baa-bd94-8edba3a82c21 ## Banks Block Meta's Muse Major U.S. financial institutions, including Charles Schwab and JPMorgan Chase, are actively blocking Meta's Muse AI from accessing client accounts. Executives emphasize that allowing third-party AI agents to move money creates uncontrollable security risks. These banks prioritize protecting client assets over the convenience of integrating external autonomous agents into their core platforms. ## Liability and Control Concerns Charles Schwab CEO Rick Wurster noted that the firm cannot control actions taken by Muse on behalf of clients. If an AI agent executes an unauthorized transaction, the question of liability remains unresolved. Consequently, banks are prohibiting these agents to ensure that human oversight remains central to all sensitive financial account movements. ## Low Adoption Rates Observed Despite the hype surrounding AI, data indicates that very few users are attempting to grant Muse access to their financial accounts. Schwab reports that blocked agent attempts remain in the tens or twenties per week. This suggests that while consumers enjoy AI tools, they remain cautious about delegating direct financial control to them. ## JPMorgan's Strict Security Stance JPMorgan Chase CEO Jamie Dimon confirmed that the bank bans all external AI agents for retail clients due to inherent risks. Dimon warned against the danger of rogue agents, noting that allowing outside software to penetrate internal systems is unacceptable. The bank maintains a defensive posture to protect its infrastructure and client data. ## In-House AI Alternatives Schwab is promoting its own in-house AI assistant, Charley, which launched last month. Unlike Muse, Charley is restricted to administrative and informational tasks, explicitly excluding trading capabilities. Wurster stated that while AI can assist with stock research, allowing autonomous decision-making without human approval represents a leap that the bank currently rejects. ## Controlled External LLM Access Schwab is preparing to support external LLM connections, but only through Model Context Protocol servers with strict authentication. These connections will be limited to viewing balances and conducting research rather than moving funds. This approach allows clients to utilize LLMs while keeping the bank’s core transaction systems isolated from external agent interference. ## API-Based Trading Limitations A subset of Schwab clients already trades through LLMs and APIs, but these interactions are strictly verified. All requests must pass through the firm's controlled APIs and MCP servers, ensuring Schwab maintains oversight. By enforcing these technical boundaries, the firm prevents external agents from bypassing security protocols or executing unauthorized financial trades. ## Strategic Defense Over Offense The banking sector's resistance highlights a broader tension between innovation and security. While AI offers potential for research, the risks of data exploitation and unauthorized access force banks to prioritize defense. Financial institutions are choosing to limit the autonomy of external agents to maintain the integrity of their systems against evolving digital threats. ## Sources covering this story - https://gizmodo.com/after-muse-launch-u-s-banks-are-saying-hell-no-to-ai-agents-moving-money-2000827221 — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._