Florida’s Attorney General wants to fine AI chatbots up to $50,000 (roughly £38,000) a time for pretending to be human, simulating a relationship, or generating sexual content without checking a user’s age first. The proposal, called the Safe Screens Act, was unveiled by Attorney General James Uthmeier at a Tallahassee press conference on 22 October 2026 and handed to state lawmakers ahead of the 2027 session. It isn’t law yet, and no bill number has been filed — but it’s the toughest state move against AI “companion” chatbots so far, and it lands squarely on a tool a lot of OnlyFans agencies already use.
Why it matters to creators
If your agency — or you — use software that flirts, sexts or chats with subscribers so a human doesn’t have to, you’re already watching a growing pile of US state law that targets exactly that. We covered the first wave of this in detail: five states already require “companion” chatbots to disclose they’re not human, with penalties up to $1,000 per message. Florida’s proposal goes further than any of them. It wouldn’t just require a disclosure banner — it would ban a chatbot outright from claiming to be human, from simulating an intimate or companionship-style relationship, and from generating sexually explicit content for anyone whose age hasn’t been verified. That is a direct description of how a lot of AI “chatter” software is built and marketed.
What the Safe Screens Act actually says
Most of the bill is aimed at protecting under-18s on social media generally: age checks at the app-store level, a curfew blocking access from 11pm to 7am, hidden like-and-comment counts, autoplay and infinite scroll switched off by default, and a two-hour daily screen-time cap. But the chatbot section stands on its own. Under the proposal, as reported by Click Orlando, any AI chatbot operator would have to determine a user’s age, and would be barred from letting an unverified or underage user access a bot that “simulates companionship through personification or emotional engagement,” suggests it’s a real person, or generates or describes sexually explicit conduct. Uthmeier framed it as closing a gap his office says Big Tech has exploited: “Parents deserve real control, and companies must design for safety, not addiction,” he said at the announcement, which Biometric Update reported in detail. Violations would be treated as unfair trade practices under Florida law, with fines up to $50,000 each — far above the $1,000-per-message private lawsuits allowed under Oregon and California’s narrower companion-chatbot laws.
The part that should worry agencies using AI chat
There’s no carve-out in the reporting so far for adult platforms or 18+-verified subscriber bases, and that’s the practical problem. A chatbot that can’t tell in advance whether the person messaging it is a verified adult has two options under a law like this: build real age verification into the chat flow, or strip out anything that could be read as “simulating companionship” or sexual content for everyone, adult subscribers included. Neither is how most of this software currently works. One widely used tool in the agency market, FlirtFlow, was marketed around building “authentic connections” with fans — language that sits uncomfortably close to what Florida wants to ban outright, not just disclose. And remember, OnlyFans’ own terms already prohibit AI-authored direct messages, so this would be a second, separate layer of exposure stacked on a practice that’s already a platform-suspension risk on your account, not the software vendor’s.
Where this actually stands
Keep this in proportion: the Safe Screens Act is a proposal Uthmeier is pushing lawmakers to adopt, not a filed, numbered bill working its way through committee. Florida’s track record on this front is real, though — its under-14 social media ban has already survived an appeals court challenge, which is partly why Uthmeier is pitching this as the next step rather than a long shot. Polk County Sheriff Grady Judd has publicly backed the plan, which tends to carry weight in Tallahassee. Nothing here is enacted, and nobody can tell you yet whether the chatbot language will survive the drafting process intact. But the direction of travel across US states is now unmistakable, and Florida just set a much higher bar than the states already on your radar.
What to do now
- Ask your agency, in writing, whether any part of your DM handling uses AI-generated text — drafted or sent — and specifically whether it’s ever presented as coming from a real person.
- If a meaningful share of your subscribers are in Florida, treat “no chatbot claiming to be human” as a live business risk, not a 2027 problem.
- Review your management contract for who carries liability for third-party chat software — if it doesn’t clearly say “the agency,” get that fixed before you sign anything new.
- Watch for an actual bill number once Florida’s legislature reconvenes — a filed bill with assigned committees is a very different thing from a press conference proposal, and the chatbot definition could still change.
- Remember this sits on top of, not instead of, OnlyFans’ existing ban on AI-written DMs and the five-state disclosure laws already in force.
This is general information, not legal advice — speak to a qualified lawyer about how these laws and proposals apply to your specific business and subscriber base.
If you’re trying to work out whether an agency’s AI chat tool or a “digital twin” licensing deal is worth the risk, we’ve covered what to check before signing one here, and our explainer on the ongoing OnlyFans “chatters” lawsuit covers the human side of the same problem — subscribers who allege they didn’t know they were talking to staff, not the creator. For a sense of why creators reach for any help with the message load in the first place, one of ours wrote about the volume of fan contact that comes with a fast-growing account.
