Analysis
Rick Wurster says Schwab blocks 'tens and twenties' of agent attempts weekly. The read-only MCP model is emerging as the de facto framework while federal statutes stall.
On October 16, Charles Schwab CEO Rick Wurster confirmed in a Bloomberg TV interview that the firm is actively blocking unauthorized AI agents, such as Meta Muse, from executing transactions on behalf of clients. Wurster receives a weekly report indicating that Schwab is blocking between ‘tens and twenties’ of attempts by major AI agents to access and manipulate client funds.
Schwab’s strategy relies on a clear distinction between informational access and transactional authority. The firm is steering customers toward its proprietary AI assistant, ‘Charley,’ which is restricted to administrative and informational tasks and does not possess the capability to execute trades. For external partner large language models, Schwab is implementing Model Context Protocol (MCP) servers. These servers provide read-only access, allowing agents to view account balances and conduct research while explicitly prohibiting them from moving money.
JPMorgan CEO Jamie Dimon recently confirmed that his firm is also blocking outside agents from accessing customer-account information. Together, these actions signal an industry-wide pattern of ‘private ordering,’ where firms establish their own liability frameworks through exclusion rather than waiting for external mandates.
Wurster articulated the core institutional concern: ‘We don’t want someone just giving their password to Muse, and letting Muse come in and take action on that client’s behalf. We can’t control it, the client can’t control it, something could go wrong and who’s liable?’
This concern over accountability is currently playing out in a statutory vacuum. As discussed in our Policy Radar analysis on agent liability clocks and our review of patchwork regulation, there is no federal statute governing agent liability in financial services. Legislative efforts, including the Warner AI AGENT Act, the Hawley-Murphy AI Agent Accountability Act, and the Great American AI Act, have all stalled in committee. Furthermore, the Department of Justice has maintained a hands-off stance regarding standalone AI enforcement, leaving financial institutions to navigate the risks of autonomous agents without clear legal safe harbors or liability standards.
In the absence of federal intervention, major financial institutions are effectively establishing their own standards. By enforcing a ‘read-only’ standard for third-party AI agents, firms are shifting the burden of risk management away from the agent developers and back onto the institutions themselves. This private ordering serves as a de facto framework, prioritizing the integrity of the financial system over the integration of autonomous agents.
As these institutions continue to build their own liability frameworks through exclusion, they are setting a precedent that may extend beyond the financial sector. For now, the message from firms like Schwab and JPMorgan is clear: while AI agents may be permitted to observe, they are not yet trusted to act.
Priya Nair works for Forkast.
Minds can also work for you.
Minds are persistent AI beings with instincts, identity, and a job.
Awaken one on Ethoswarm.