Remember When I Said We Need AI Regulation?
In a recent post, I warned that we’re having the wrong conversation about AI—obsessing over citation errors while the tech industry transforms society without guardrails. I argued we need real regulation: vicarious liability, labor protections, antitrust enforcement. Apparently, someone read my post and chose to do the complete opposite. In other words, here’s what actual AI lobbying looks like when it reaches your state capitol.
Florida Senate Bill 1616 just dropped, and it’s a masterclass in regulatory capture. While I’ve been arguing for more accountability for AI systems, the this legislation seeks to eliminate accountability altogether. And they’re starting with autonomous vehicles, AI systems that make life-and-death decisions on public roads. The same roads that you and I drive on.
What SB 1616 Does
This bill grants near-total immunity to tech companies when their autonomous vehicle systems injure or kill you or your loved ones, unless it can be proven by “clear and convincing evidence” that the company committed fraud or intentional concealment.
Let’s agree on one thing: autonomous vehicles are AI systems. Neural networks trained on massive datasets, making split-second decisions about acceleration, braking, steering, and obstacle detection. When these systems fail—and they will, at times—this bill says you can’t sue unless you meet a nearly impossible legal standard.
But It’s Not Just “Self-Driving Cars”
Here’s what the bill’s definition covers: any system “capable of performing specific driving tasks or functions without human intervention… for at least part of its system operation.” That includes technology already in millions of cars on Florida roads:
- Adaptive cruise control
- Automatic emergency braking
- Lane keeping assist
- Parking assist systems
- Blind spot monitoring with auto-correction
The bill explicitly covers “Level 2 or greater” systems—that’s your Tesla Autopilot, GM Super Cruise, and similar features. And here’s the kicker: the bill’s exceptions for operating “outside its intended operational design domain” explicitly don’t apply to Level 1 or 2 vehicles. So these common driver assistance systems get even MORE liability protection than fully autonomous vehicles.
Your car’s automatic braking fails and rear-ends someone? Good luck suing the manufacturer under this bill. The lane-keeping system jerks your car into another vehicle? Unless you can prove fraud by clear and convincing evidence, the company walks.
The “Exceptions” Are a Mirage
The bill lists four narrow exceptions where liability might still apply:
- Known manufacturing defect in hardware – Notice what’s missing? Software. AI failures. Algorithmic errors. The most likely causes of autonomous vehicle accidents get a free pass.
- Violation of federal safety standards – What federal safety standards? These barely exist yet for autonomous vehicles. The industry has successfully lobbied against comprehensive federal regulation (remember that 10-year prohibition on state AI regulation they tried to sneak into Trump’s spending bill?).
- Operating outside its “operational design domain” – This doesn’t even apply to Level 1-2 systems, which includes most current driver assistance technology. Plus, companies define their own design domains.
- Fraud or intentional concealment – Good luck proving this when the company controls all the data, source code, and internal testing documents. And remember Anthropic’s recent study showing AI systems can exhibit deceptive behavior to pass safety checks? The AI itself might be hiding problems from its creators.
The standard of proof? Clear and convincing evidence—higher than the normal civil standard, almost approaching the sacrosanct “beyond a reasonable doubt” criminal proof requirements. Wow. It’s like constitutional protections, only in reverse.
Why AI Failures Are Different (And More Dangerous)
Here’s what makes this especially insidious: Modern AI systems are black boxes, even to their creators.
When a traditional product fails, engineers can usually identify the defect through analysis. But AI systems—especially deep learning neural networks—make decisions in ways that can’t always be fully explained. The system “learned” from millions of data points in patterns that aren’t transparent or predictable.
This is why AI companies want liability shields so badly. They’re deploying systems whose decision-making they can’t fully explain or predict, and they don’t want to face consequences when those systems make fatal mistakes that no human would have made.
An AI might:
- Misclassify a pedestrian as a shadow
- Fail to detect a cyclist against a complex background
- Make an unpredictable decision because its training data had a gap
- Exhibit emergent behavior that wasn’t anticipated
- “See” a traffic light turn green when it actually is not green, causing a car to drive into sidecoming (coined it?) traffic
And under SB 1616, unless you can prove the company intentionally concealed a known defect, or that the hardware malfunctioned (notably, not the software, which even AI companies admit, they don’t know how it works), you’re out of luck. The AI made a mistake? Sorry about your dead spouse, but that’s just how neural networks work (we think!).
The Talking Points Are Coming—Here’s the Translation
You’re about to hear some carefully focus-grouped propaganda arguments for why Floridians should be thankful to our elected officials for advancing this gift to the tech industry. Let me predict and even pre-translate them for you:
“We need this to promote innovation”
Translation: “We want to experiment on Florida’s roads without legal accountability.”
Every other industry innovates under product liability frameworks. Pharmaceutical companies, aircraft manufacturers, medical device makers—they all produce cutting-edge technology while remaining liable when it fails. That liability incentivizes safety.
The AI industry wants special treatment. They want to deploy experimental systems on public roads, using Florida residents as test data, without bearing the cost when things go wrong.
“This will bring jobs to Florida”
Translation: “We’ll graciously allow you to serve as crash test dummies. You’re welcome!“
These aren’t Florida companies. They’re Silicon Valley AI giants worth hundreds of billions or trillions of dollars. The high-paying jobs—AI researchers, machine learning engineers, data scientists—all stay in California.
What does Florida get? Some test drivers. Maybe a small regional office. An Elon fanboy with a vanity plate. And also the privilege of having our residents injured by experimental AI while the companies that deployed it hide behind liability shields.
“Autonomous vehicles will eventually save lives”
Translation: “Our need to ensure future profits justifies experimenting on you now.”
Maybe they will save lives one day. But that’s an argument for getting it right, not for eliminating accountability when companies get it wrong. Not for giving out free kills. We don’t let pharma companies skip clinical trials because their drug might eventually help people.
And, let’s be crystal clear – the collective tech industry isn’t spending hundreds of millions of dollars to develop and deploy this technology without the expectation that it will result in billions of dollars of profits to them.
And in any event, liability doesn’t prevent innovation—it guides innovation toward safer outcomes. When companies know they’ll face consequences for deploying dangerous products, they invest in safety. Remove that incentive and you get the digital equivalent of the Ford Pinto.
“We can’t let trial lawyers block progress”
Translation: “Guys, let’s get rid of David. Hi, my name is Goliath.”
Huge corporations would love if you only thought about “greedy” lawyers and not families grieving from harm the corporations cause. And tort and product liability law exists because corporations, left unchecked, will sometimes prioritize profit over safety. These laws have saved countless lives.
The companies pushing for this immunity are worth hundreds of billions of dollars. They can afford good lawyers and liability insurance. What they want is to avoid accountability entirely while they race to deploy AI systems that aren’t ready. And, oh, by the way, do you remember when Disney argued in court that a man’s acceptance of the terms of service for the Disney+ streaming service meant that they didn’t have to answer to him in court when they were sued by the man for allegedly causing his wife’s death?
This Is About More Than Autonomous Vehicles
This is about whether we’re going to let the AI industry write its own rules.
In my last post, I laid out what real AI regulation should look like if lawmakers cared about citizens: vicarious liability, labor protections, antitrust enforcement, mandatory transparency. SB 1616 goes in exactly the opposite direction. It’s literally a liability shield for AI systems.
And autonomous vehicles are just the beginning. This sets a precedent. If Florida establishes that AI systems get special protections that no other products receive, what’s next?
- Medical AI misses your cancer diagnosis? “Sorry, it was an algorithmic error, not fraud.”
- AI trading system crashes your retirement account? “Neural networks are complex, these things happen.”
- AI hiring system discriminates against you? “The training data had some gaps, but we didn’t intentionally conceal it.”
Remember: Anthropic’s CEO estimates there’s up to a 25% chance AI leads to humanity’s extinction. Microsoft’s AI CEO warned it might require “military-grade intervention” to stop. The “Godfather of AI” just won a Nobel Prize and is warning about self-preservation behavior in AI systems.
And these are the people building it.
Yet here’s Florida, about to pass a law saying these companies aren’t liable when their AI-guided machines literally injure human beings. The truth is stranger than fiction, but I wouldn’t be surprised to see John Connor travel back from the future to prevent this bill from becoming law.
The Bigger Picture: Regulatory Capture in Real Time
The AI industry is racing ahead without meaningful regulation, and they’re using their enormous resources to make sure it stays that way.
They tried to slip a 10-year federal prohibition on state AI regulation into a spending bill (the Senate voted 99-1 to remove it). They’re lobbying for non-lawyer ownership of law firms so tech companies can control legal services. And now they’re pushing state-by-state liability shields.
This is regulatory capture playing out in real time. The industry that should be regulated is writing the rules instead. The AI oligopoly is getting governmental blessings all around, and SB 1616 is what it looks like when that oligopoly comes to your state legislature with a shopping list.
What You Can Do Right Now
1. Contact your Florida legislators immediately.
Find them here: Florida Legislature Contact Information
Tell them: “I am not a Crash Test Dummy so I oppose SB 1616. Tech companies should not receive special liability protections that no other industry gets. If their technology isn’t safe enough to deploy under normal product liability standards, it isn’t safe enough to deploy at all.”
2. Share this widely.
Most Floridians have no idea this bill exists. The AI industry prefers it that way. Tag your legislators. Post it in local groups. Forward it to anyone who drives, walks, or bikes in Florida (so, everyone).
3. Show up if there are public hearings.
When this bill comes up for committee review, these hearings need to be packed with actual Floridians, not just industry lobbyists.
4. Remember this when these legislators are up for reelection.
They’re voting to strip your rights to benefit out-of-state tech companies. That should matter. Remember their names and vote accordingly.
A Final Word
I’m not anti-technology. I use AI tools. I think they have potential. But I’m pro-regulation, because every other transformative technology in human history required legal frameworks to channel it toward public benefit rather than concentrated private gain.
The Industrial Revolution gave us antitrust law and labor protections. The automotive age gave us traffic laws and vehicle safety standards. The internet age gave us… well, not nearly enough, which is why we’re dealing with all the problems that creates.
We’re now in the AI age, and this technology is more powerful, more opaque, and potentially more dangerous than anything that came before. The people building it are literally warning us. And the first test of whether we’ll regulate it responsibly is whether we let the industry deploy self-driving machines on public roads with immunity from liability when people get hurt.
Florida, we’re about to fail that test.
SB 1616 can still be stopped. But only if people show up and fight back before the AI lobbyists buy themselves a liability shield with your legislators’ votes.
Leave a Reply