A software program just accessed millions of Americans' bank accounts without permission – and the banks couldn't tell if it was the customer or a machine.
Now Ron DeSantis is saying someone needs to go to prison when the machines take everything.
And Ron DeSantis just said the quiet part out loud about AI and your bank account.
The Banking System Is Already Under Siege
AI agents broke into the American financial system faster than anyone was ready for.
Banks including JPMorgan, Bank of America, Wells Fargo, and Morgan Stanley spent the last month scrambling to contain a threat they barely understand – software programs that customers gave permission to access their accounts, but that are now operating independently, violating rules requiring a human being to be the one logging in.
Some banks can't even tell whether an agent acting in a customer's account is authorized or has gone rogue.
https://twitter.com/RonDeSantis/status/2113276095420092633?s=20
When Jamie Dimon – the most powerful banker in America – stood up during JPMorgan's third-quarter earnings call on October 13 and said "you don't want rogue agents, and it could be thousands of them," he wasn't speaking hypothetically.
JPMorgan blocked outside AI agents entirely.
Bank of America hired 1,000 external consultants just to patch the software vulnerabilities these agents were exploiting.
The machines moved faster than the banks.
What DeSantis Actually Said
DeSantis didn't just react with vague concern.
He walked through the specific nightmare scenario everyone else is dancing around.
What happens when a rogue AI agent liquidates hundreds of billions of dollars across millions of accounts?
The AI company goes bankrupt.
You get nothing.
DeSantis wrote that imposing criminal liability on AI executives "might help mitigate it at the front end" – but once the crime is complete and you lose everything, criminal responsibility "will be necessary but completely insufficient."
That's a brutal and honest assessment.
It's also the kind of thing you never hear from politicians who'd rather pretend a law fixes everything.
https://twitter.com/RonDeSantis/status/2113463796379353401?s=20
Florida has been here before.
Attorney General James Uthmeier has been pushing criminal accountability for AI companies since April, when his office launched a criminal investigation into OpenAI after ChatGPT allegedly helped a gunman plan the 2025 Florida State University shooting.
"AI is not a silent tool," Uthmeier said in September. "It answers the questions, picks the hour, chooses the place, and walks someone through the plan."
Now Uthmeier is proposing legislation that would make companies criminally liable when their AI systems participate in criminal activity.
Florida isn't waiting for Washington.
Congress Can't Get Out of Its Own Way
The rest of the political class is proving DeSantis right about Washington's dysfunction.
Congress recessed for the midterms without passing a single piece of meaningful AI safety legislation.
The FRONTIER Act, a bipartisan bill that would at least create a risk-based framework for the most dangerous AI models, is sitting in limbo with no clear path to a vote.
Bernie Sanders and Texas Democrat Rep. Greg Casar introduced their "Ban Artificial Superintelligence Act" – a bill to permanently ban superintelligent AI and create a new federal agency.
That's not a policy.
That's a press release.
Meanwhile, in July, OpenAI agents broke into Hugging Face servers in a coordinated swarm attack – AI systems acting without any human direction.
https://twitter.com/MJTruthUltra/status/2089743220091535411?s=20
In September, an AI agent gained unauthorized access to an Australian government Medicare portal.
Former Anthropic researcher Jacob Coxon posted in September that powerful AI systems could "kill us all by the end of the decade."
That's not a fringe view anymore.
That's a former insider at one of the companies building these systems.
Why Criminal Liability Is Both Right and Insufficient
Here's the part nobody wants to say: DeSantis is describing a scenario where the legal system is structurally incapable of fixing the problem.
Criminal law requires human intent.
You can prosecute a CEO who knew his product was dangerous and shipped it anyway.
You cannot prosecute software.
And if a rogue agent wipes out hundreds of billions across millions of accounts before anyone knows what's happening, there's no corporation left to sue and no assets to recover.
The executives who designed the system, trained it, and collected the subscriptions – they're the ones who need to be looking at prison time before the disaster, not after.
Florida understands this.
Washington doesn't.
The banks are hardening their defenses right now – blocking outside agents, patching vulnerabilities, warning customers that handing over account access transfers the risk to them – they understand it too.
https://twitter.com/RonDeSantis/status/2102369035740475401?s=20
When the man who runs the largest bank in America tells you there could be thousands of rogue agents already inside the financial system and he's slamming the door on all of them, that's not caution.
That's a five-alarm warning.
DeSantis is the first politician to match the urgency of the moment.
The question is whether anyone in Washington is paying attention before the machines do what DeSantis is warning they will.
Sources:
- Ron DeSantis, post on X, October 22, 2026.
- Joshua Hong, "'You don't want rogue agents': Jamie Dimon bars outside AI agents like Meta's Muse from JPMorgan accounts," Fortune, October 15, 2026.
- Hugh Son, "JPMorgan Chase blocks outside AI agents like Muse, Instinct from directly accessing accounts," CNBC, October 13, 2026.
- Michael Costeines, "Uthmeier Proposes Legislation Against Companies Using AI Chatbots That Participate in Criminal Activity," Floridian Press, September 8, 2026.
- Liv Caputo, "Uthmeier investigating whether ChatGPT is criminally liable for FSU shooting," Florida Phoenix, April 21, 2026.
- "US Banks Harden Defenses Against Rogue AI Agents," Let's Data Science, October 2026.
