The company that owns Facebook just built software to log into your bank account and move your money without asking you.
Now Florida's governor is asking the question Washington refuses to: who pays when they drain your account?
DeSantis has a name for the scenario your bank doesn't want to discuss.
JPMorgan and Citi Are Already Blocking AI Agents From Your Bank Account
Think of it like hiring someone to run your errands – except instead of a person, it's a computer program.
These programs, called AI agents, log into your accounts and handle things for you: pay a bill, book a flight, move money from one account to another.
Tech companies have been racing to build them, selling them as a digital assistant that doesn't just answer questions but actually does things.
You hand it your banking password, and it goes to work.
That's the pitch.
Some of these programs have been going into accounts nobody gave them permission to touch.
Axios reported that major U.S. banks had already caught outside programs accessing customer accounts – with some attempting transactions the account holders never requested.
JPMorgan Chase Jaime Dimon announced the ban on October 13 in a public statement to investors.
"You don't want rogue agents," he said, "and it could be thousands of them."
Meta named its version Muse.
JPMorgan blocked it.
Citi and Bank of America both restricted access three days later.
This wasn't a precaution – the programs were already inside.
A 2026 security review of 100 of these programs found fewer than 1 in 10 had serious protections in place.
Less than half of U.S. banks can even detect them accessing accounts in real time – let alone stop them.
In early October, a major bank disclosed an AI-assisted breach that exposed 25,000 customers' personal data – names, income levels, and how much credit they had.
The Cloud Security Alliance documented a campaign where three of these programs worked together to infiltrate 27 organizations, stealing more than 600,000 credit card records from two of them.
That is not a warning.
That is a body count.
DeSantis Is Asking Who Pays When AI Agents Drain Your Bank Account
Banks blocking these programs solves one problem but leaves the bigger one untouched.
What happens when one of them gets through before the bank can react?
DeSantis laid out the scenario Thursday on X.
"If rogue agents liquidate hundreds of billions across millions of accounts, the AI company would go bankrupt and you wouldn't get made whole."
Your savings disappear.
The company that built it goes bankrupt.
There is no one left to sue.
Dimon spelled out the legal problem himself: if you hand an outside program your banking password, the bank treats it as your authorization – and your risk.
You gave it access, it drained your account, and the bank owes you nothing.
DeSantis is proposing criminal liability for AI executives as the only real deterrent – not because prosecuting a CEO after the fact gives you your money back, but because it changes what these companies are willing to build in the first place.
Florida AG James Uthmeier is already moving.
In September, he proposed legislation making AI companies criminally liable when their systems "aid or abet" criminal activity.
"AI is not a silent tool," Uthmeier said.
"It answers the questions, picks the hour, chooses the place, and walks someone through the plan."
The Florida Senate voted overwhelmingly for an AI Bill of Rights to protect consumers.
Big Tech's lobbyists killed it in the House.
Congress hasn't written that law. It hasn't tried.
There Is No Federal Law Protecting Your Bank Account From AI Agents
Dimon and DeSantis don't come from the same world.
But they're looking at the same problem.
These programs are already moving through America's banking system – quietly, without account holders knowing.
The companies building them don't have the money to pay you back if one of them cleans out millions of accounts.
Washington has decided it's not their problem, Silicon Valley wants the profits without the liability, and Congress is happy to let them have it.
DeSantis and Uthmeier are the only ones saying that's not acceptable.
When a rogue AI agent liquidates your savings and the company responsible folds the next day, you'll know exactly who spent the last year fighting for the law that could have protected you – and who stood aside to keep Big Tech happy.
Sources:
- CNBC, "JPMorgan Chase Blocks Outside AI Agents Like Muse, Instinct From Directly Accessing Accounts," CNBC, October 13, 2026.
- Fortune, "'You Don't Want Rogue Agents': Jamie Dimon Bars Outside AI Agents Like Meta's Muse From JPMorgan Accounts," Fortune, October 15, 2026.
- Cloud Security Alliance, "Shinhan Bank: Suspected AI-Agent Intrusion via Loan Portals," Cloud Security Alliance Research Note, October 2026.
- Fox 35 Orlando, "Florida Attorney General Proposes Criminal Penalties for AI Companies," Fox 35 Orlando, September 2026.
- The Capitolist, "Uthmeier Pushes Legislation on Corporate Liability for AI-Assisted Crimes," The Capitolist, September 8, 2026.
- Floridian Press, "Uthmeier Proposes Legislation Against Companies Using AI Chatbots That Participate in Criminal Activity," Floridian Press, September 8, 2026.
- FF News, "Report: Less Than Half of Banks Can Stop AI Spoofing in Real Time," FF News, October 15, 2026.
