Florida Sues ChatGPT: Should AI Have Guardrails? | Wine by the Case
Artificial intelligence has officially entered its lawsuit era.
In this episode of Wine by the Case, the conversation turns to one of the biggest questions surrounding AI right now: Who is responsible when artificial intelligence goes wrong, and how many guardrails should companies be required to put in place?
Florida Attorney General James Uthmeier filed a first-of-its-kind state lawsuit against OpenAI and CEO Sam Altman on June 1, 2026. The lawsuit alleges that OpenAI aggressively marketed ChatGPT while failing to adequately disclose or address potential risks, particularly those involving children, privacy, mental health, and harmful uses of the technology. OpenAI disputes the allegations and says it continues to strengthen its safety systems.
And because this is Wine by the Case, we are pairing a complicated legal question with a California blend that is almost as unconventional as the conversation.
When AI Stops Being “Just a Search Engine”
One of the central issues raised in this episode is the difference between asking Google a question and having an ongoing conversation with an AI chatbot.
ChatGPT doesn't simply return a page of search results. It generates responses based on the conversation happening in real time. That distinction has become increasingly important as lawmakers, parents, courts, and technology companies wrestle with questions about how people interact with AI.
Florida's lawsuit argues that OpenAI has not done enough to address risks associated with those interactions. Among other claims, the state alleges problems involving protections for children, the collection of minors' data, product safety, and representations made about ChatGPT. The complaint includes claims involving negligence and Florida's laws against deceptive and unfair trade practices.
Those are allegations, not findings of liability, and the litigation is ongoing.
So, Should ChatGPT Have More Guardrails?
That sounds like a simple question.
It isn't.
AI systems already have safety measures, but deciding where those boundaries belong raises another set of questions.
How much responsibility belongs to the company creating the technology? How much belongs to the person using it? Should different protections apply to children? And when does protecting users cross over into limiting legitimate uses of a general-purpose tool?
OpenAI says ChatGPT includes safeguards designed to recognize sensitive situations, respond appropriately, and direct users toward real-world support when necessary. The company has also pointed to parental controls and other protections for younger users.
Florida argues those protections have not gone far enough.
That disagreement is now playing out in court.
What Happens to Your ChatGPT Conversations?
Then we get to the question that should make every lawyer lean a little closer to the microphone:
Are your AI conversations private?
The episode looks at privacy from several angles, including investigations, litigation, and the possibility that AI conversations could become relevant evidence in a lawsuit.
Just because a conversation feels private does not necessarily mean it is legally protected by attorney-client privilege or another recognized privilege. If an AI chat becomes relevant to litigation, questions about preservation, subpoenas, discovery, account records, and admissibility could become significant.
In other words, typing something into an AI chatbot is not necessarily the digital equivalent of whispering it into a locked diary.
That makes the increasingly popular habit of sharing incredibly personal details with AI worth examining from both a privacy and legal perspective.
And Then There’s “Unhinged ChatGPT”
Naturally, we also had to discuss the viral trend asking ChatGPT to describe its users in increasingly outrageous ways.
Because nothing says responsible AI governance quite like asking a computer to roast you.
It raises a surprisingly useful point, though. People increasingly treat AI as something more conversational than traditional software. We joke with it. Ask it for advice. Give it context. Tell it things we would never type into a basic search box.
And the more personal those interactions become, the more important questions about privacy, safety, transparency, and responsibility become.
The Featured Wine: Stolpman Vineyards Syrah-Sangiovese
The legal discussion may be complicated, but this episode's featured wine keeps things considerably more straightforward.
From California's Ballard Canyon AVA, the Stolpman Vineyards blend combines 50% Syrah and 50% Sangiovese.
The Syrah contributes richness and structure while the Sangiovese brings brighter acidity and elegance. Expect dark berry, black cherry and plum flavors alongside pepper, earthy spice and a long finish.
Wine Notes
Region: Ballard Canyon, California
Blend: 50% Syrah / 50% Sangiovese
Body: Full-bodied
Tasting Notes: Blackberry, black cherry, plum, pepper and earthy spice
Finish: Rich and lingering with balanced acidity
Pairings: Grilled steak, lamb, pasta with red sauce and charcuterie
Wine by the Case Rating: 8.5/10
AI, Privacy and the Law Are Only Getting More Complicated
Florida's lawsuit could become an important test of how existing consumer protection and product liability laws apply to rapidly evolving AI systems.
As of August 2026, the case is still in its early stages. OpenAI has sought to move the case into federal court, while Florida has argued that it belongs in state court. The underlying allegations have not been decided.
Whatever ultimately happens, the larger questions aren't going away.
How safe should AI have to be? What responsibilities should companies have toward younger users? How private are our conversations? And how will courts treat the enormous amount of information people are voluntarily handing over to AI systems?
Pour into the latest episode of Wine by the Case for the full conversation.
Watch the episode: Florida AG Sues ChatGPT & Sam Altman: Should AI Have Guardrails? | Wine by the Case
Wine content is intended for audiences of legal drinking age.