AI and Copyright Law in 2026: What Florida Businesses Need to Know

By the time you finish reading this sentence, somewhere in Florida a marketing team has generated a new product description with ChatGPT, a designer has prompted Midjourney for a logo concept, and a developer has asked Claude or Copilot to write a block of code. Generative AI has moved from novelty to infrastructure — and the legal questions it raises are no longer theoretical.

Who owns the output of an AI system? Can a Florida business safely use AI-generated images in advertising? What happens when the AI was trained on someone else’s copyrighted work? And with Florida now moving its own AI Bill of Rights through the legislature, what rules apply here specifically?

This guide walks Florida business owners, SaaS founders, creators, and in-house counsel through the 2026 state of AI copyright law, the court rulings that have reshaped the field in the last eighteen months, and the practical steps you should take to protect your intellectual property.

Key Takeaways

• AI cannot hold a copyright. Only humans can be authors under U.S. law. The Supreme Court confirmed this in March 2026 by declining to hear the appeal in Thaler v. Perlmutter.
• AI-assisted works can be protected — but only if the human contribution is creatively substantial. Typing a prompt is not enough.
• Training data lawsuits are reshaping the industry. The $1.5 billion Bartz v. Anthropic settlement and the Thomson Reuters v. Ross Intelligence appeal have set new expectations for how AI companies source data.
• Florida is regulating AI directly. The state’s AI Bill of Rights (SB 482) adds consent requirements around name, image, and likeness, and the Florida Digital Bill of Rights (FDBR) governs how personal data is used in AI systems.
• Businesses need a documented AI use policy. Copyright ownership, client deliverables, and vendor indemnities all depend on it.

Why This Matters Right Now

Three things changed the landscape between 2024 and 2026: the U.S. Copyright Office published its comprehensive three-part report on AI and copyrightability, federal courts issued the first substantive rulings on AI training data, and Florida began building its own AI-specific legal framework.

For a Tampa SaaS company embedding AI features into its product, a Florida agency using generative tools for client deliverables, or an inventor drafting a patent application with AI help, these developments are not abstract. They determine whether your work is protectable, whether your vendor agreements are enforceable, and whether your AI use exposes you to infringement claims.

Can AI-Generated Work Be Copyrighted?

The short answer is no — not on its own. U.S. copyright law requires a human author. This has been the position of the U.S. Copyright Office since 2023, and the D.C. Circuit Court of Appeals affirmed it in Thaler v. Perlmutter in March 2025. In March 2026, the Supreme Court denied certiorari, making the rule settled law at every level of the federal judiciary.

But the more practical question is: can work you created with AI be copyrighted? Here the answer is a qualified yes.

The U.S. Copyright Office distinguishes between:

1. AI-generated works — output produced by an AI system with no meaningful human creative input. Not protectable.
2. AI-assisted works — output where a human exercised substantial creative control over the expressive elements. Potentially protectable, but only the human-authored portions are covered.

The line between the two is fact-specific. The Copyright Office has been clear that entering a text prompt — even a detailed one — does not make the user an author of the resulting output. What matters is whether the human made the meaningful creative decisions about the final expression: selection, arrangement, editing, refinement, and integration of AI output into a larger creative work.

A photographer who uses an AI tool to remove a distracting background from their own photograph retains copyright in the photograph. A novelist who drafts a scene with AI and then rewrites, restructures, and integrates it into their manuscript owns the novel. A marketer who generates fifty logo options with Midjourney and submits the raw output without modification likely owns nothing.

Practical checklist for Florida businesses

If your business relies on AI-assisted creative output, document:

• The specific human creative decisions made at each stage
• Which elements were AI-generated and which were human-authored or human-modified
• The prompts, iterations, and refinement steps
• Who on your team contributed what

This documentation is what supports a later copyright registration — and what protects you if someone challenges your ownership.

The Training Data Question: Fair Use in 2026

If AI output ownership is one side of the coin, AI training data is the other. The central legal question — whether it is lawful to train a generative AI model on copyrighted works without permission — is working its way through the federal courts in a cluster of high-stakes cases.

Thomson Reuters v. Ross Intelligence

In early 2025, a federal court granted summary judgment to Thomson Reuters, ruling that Ross Intelligence’s use of Westlaw headnotes to train an AI-driven legal research tool was not fair use. The case is now on appeal before the Third Circuit. Notably, Ross Intelligence was not a generative AI company — its product was an AI-powered case search tool — which has led commentators to question how broadly the ruling applies to generative models.

Bartz v. Anthropic

A class of authors sued Anthropic over the use of their books in training data. The district court reached a split ruling: training an AI model on lawfully acquired copyrighted books was fair use, but storing pirated copies of those books was not. Anthropic settled the pirated-copies portion of the case for $1.5 billion, with payouts estimated at roughly $3,000 per work. The opt-out window for the settlement closed in February 2026, and final court approval is expected later this year.

The OpenAI MDL

A consolidated multidistrict litigation against OpenAI continues in federal court. In January 2026, the court ordered OpenAI to produce all 20 million output logs relevant to the plaintiffs’ claims — a significant discovery ruling. The case centers on allegations that OpenAI’s models were trained on copyrighted books and that ChatGPT can be prompted to produce infringing summaries and outlines of those works.

What this means for Florida businesses

Two practical consequences flow from the current state of training-data litigation:

1. If you license AI tools, your vendor contracts should include robust indemnities for copyright infringement claims arising from the training data. Most major providers now offer some form of indemnity, but coverage varies significantly — read the fine print.
2. If you build AI tools, your own training data sourcing is now a legal risk you must document and manage. The “we’ll figure it out later” posture that characterized 2022–2023 is no longer defensible.

Florida’s AI Bill of Rights and the FDBR

While federal copyright law governs ownership, Florida has taken an active role in regulating how AI can be used — particularly when it touches personal data, minors, or a person’s name, image, and likeness.

The Florida AI Bill of Rights (SB 482)

Filed by Senator Tom Leek and advancing through the 2026 legislative session, Florida’s AI Bill of Rights establishes a set of consumer-facing protections:

• The right to know whether you are communicating with a human or an AI system
• Restrictions on the commercial use of a person’s name, image, or likeness generated through AI without consent
• Consent requirements for minors using AI companion chatbots, with additional disclosure and protection obligations on the platforms
• A prohibition on the sale or disclosure of personal information by AI technology companies unless the data is deidentified
• Private causes of action and civil penalties for violations

For Florida businesses, the name-image-likeness provisions matter most in marketing, advertising, entertainment, sports, and influencer contexts. If your agency generates AI imagery depicting a real person, the consent rules under SB 482 — combined with existing right-of-publicity doctrine — create real liability exposure.

The Florida Digital Bill of Rights (FDBR)

Effective July 2024, the FDBR is Florida’s comprehensive data privacy law. It applies to businesses that conduct commerce in Florida or process personal data of Florida residents above certain thresholds. The FDBR’s data minimization, consent, and non-discrimination requirements apply just as forcefully to AI systems as to any other data processing activity.

If your SaaS product uses customer data to fine-tune or improve an AI model, the FDBR likely governs how you obtain consent, how you describe the use in your privacy notice, and how you honor consumer rights to access, correct, or delete that data.

Political advertising disclaimers

Florida also has a specific statute — Fla. Stat. 106.145 — that requires prominent disclaimers on political advertisements containing AI-generated content depicting a real person. This applies to political committees, candidates, and agencies producing campaign material in Florida.

Common AI Copyright Scenarios — And How to Handle Them

1. Using AI-generated images in marketing

Risk: The underlying AI model may have been trained on copyrighted images; the output may incorporate protectable elements from those sources; you cannot copyright the image itself.

Best practice: Use enterprise AI tools that offer copyright indemnification. Keep records of your prompt and iteration process. Modify the output with substantial human creative input before commercial use. Avoid prompts that reference specific artists, brands, or copyrighted characters.

2. AI-assisted software development

Risk: AI coding tools trained on open-source repositories may produce output that echoes licensed code. Certain open-source licenses (GPL, AGPL) carry copyleft obligations.

Best practice: Use coding assistants that filter for license-flagged matches. Maintain a software bill of materials. Require human code review before merging AI-generated code into your production codebase. Update your contributor license agreement and open-source policy to address AI-assisted contributions.

3. AI-written marketing copy and blog content

Risk: Pure AI output cannot be copyrighted. If your content is scraped and republished, you have limited recourse.

Best practice: Treat AI as a drafting tool. Add substantial human editing, restructuring, and original analysis. Register your compiled, edited works with the U.S. Copyright Office when appropriate.

4. AI-generated product designs or logos

Risk: A raw AI-generated logo is not copyrightable. It may also be unregistrable as a trademark if it too closely resembles existing marks in the AI’s training data.

Best practice: Conduct a thorough trademark clearance search before adopting any AI-generated brand asset. Work with a designer to make substantial human modifications. File a trademark application rather than relying on copyright alone — trademark rights arise from use in commerce, not from authorship.

5. AI features embedded in your SaaS product

Risk: Your users may input copyrighted material. Your output may infringe. Your training data may create liability.

Best practice: Terms of service should allocate IP risk clearly. Provide users with indemnity only to the extent your vendor indemnifies you. Implement input and output filtering. Disclose AI use in your privacy policy and comply with FDBR consent requirements.

When to Consult an Attorney

Most Florida businesses using AI do not need custom legal advice for routine tasks. But certain situations call for a conversation with intellectual property counsel:

• You are about to commercialize an AI-assisted work and want to understand what is protectable
• You are building an AI product and need to structure training data sourcing, user agreements, and vendor contracts
• You have received a takedown notice, a cease-and-desist, or a copyright infringement claim tied to AI use
• You are negotiating an acquisition, investment, or licensing deal that depends on IP ownership
• Your agency or firm uses AI in client deliverables and you need a defensible workflow and contract language

At Luby & Rauscher PA, we combine patent, trademark, copyright, and technology law experience — grounded in Justin Luby’s engineering background and USPTO registration — to advise Florida businesses on exactly these issues.

Frequently Asked Questions

Can I copyright something I created using ChatGPT or another AI tool? Only if your human creative contribution was substantial. Typing a prompt and using the raw output will not qualify. Editing, restructuring, integrating, and refining the output into a larger human-authored work can qualify.

Is AI training on copyrighted material legal? It depends. Courts have issued inconsistent rulings, and most cases are still pending or on appeal. Training on lawfully acquired copyrighted works has received some fair-use protection; training on pirated copies has not. Expect continued litigation and, eventually, Supreme Court review.

What happens if I use an AI-generated image and it turns out to resemble copyrighted art? You could face a copyright infringement claim. Many enterprise AI vendors now offer indemnification to cover this risk, but only for paid tiers and only under specific conditions. Free tools generally offer no such protection.

Does the Florida AI Bill of Rights apply to my business? If you do business in Florida or market to Florida consumers and use AI, some provisions likely apply. The name-image-likeness consent rules, chatbot disclosure requirements, and data-sale restrictions have broad reach. The statute is still moving through the legislature, so the final scope may change.

Can I register a trademark for an AI-generated logo? Trademark rights are separate from copyright and arise from use in commerce. You can register a trademark for a logo regardless of how it was created — but you should first confirm the design is not confusingly similar to existing registered marks. A trademark clearance search is essential.

Do I need a written AI policy for my company? If your team uses AI tools for any business purpose, yes. A good AI policy covers approved tools, data handling, human review requirements, ownership of AI-assisted outputs, vendor indemnities, and compliance with Florida data privacy law.

Closing Thoughts

The law around AI and copyright is moving fast, but the core principles are becoming clearer. Human authorship is required. Documentation is protection. Vendor contracts are where most of your real risk sits. And Florida, like every state, is writing its own chapter of AI regulation in parallel with the federal courts.

If you run a Florida business and AI is part of how you create, deliver, or productize value, now is the time to get your house in order — before a takedown notice, a contract dispute, or a regulatory inquiry forces the issue.

---

Call to Action Block

Have questions about AI, copyright, or protecting your Florida business’s intellectual property?

Luby & Rauscher PA is a Tampa Bay IP and technology law firm led by Justin Luby, a USPTO-registered patent attorney with over 20 years of engineering experience. We help Florida SaaS companies, agencies, inventors, and creators navigate the evolving intersection of AI, copyright, and technology law.

Schedule a free consultation: (813) 919-6958 | info@fltechlaw.com Visit: 4326 W. El Prado Blvd, Suite 5, Tampa FL 33629