There is no single intellectual property strategy for an AI startup. The optimal approach is a portfolio strategy in which different assets—model weights, generated code, interface designs, marketing copy—are matched to different legal tools: trade secrets, copyrights, patents, and trademarks. Legal commentators writing for AI companies converge on this point, and it changes how founders should allocate scarce early-stage attention.
Attorney Philip P. Crowley of Crowley Law LLC, writing on AI-generated startup ownership, breaks the question down asset by asset. His analysis shows that the protection available for each component depends on how much human creative input went into it, whether it can be kept confidential, and who else might claim rights in it. Terri Shieh-Newton, a patent attorney writing for the firm Mintz, adds the strategic layer: deciding what to patent versus what to hold as a trade secret is among the most consequential IP decisions an AI company makes.
Mapping AI Assets to IP Protection: A Decision Framework
The core insight from Crowley's asset-by-asset analysis is that AI outputs are not automatically protected. Raw AI-generated code, artwork, and text generally start life without copyright protection, because copyright requires human authorship. What a founder does next—editing, arranging, restricting access—determines whether protection attaches. The framework below maps four common AI startup assets to their recommended protection, the primary risk each faces, and the action required to secure it.
AI Startup IP Protection Decision Framework
Founders can use this table to quickly identify the recommended protection, key risk, and necessary action for each asset class. The recommendations follow Crowley's ownership analysis for AI-generated work.
| AI Asset Type | Primary IP Protection | Key Risk | Protective Action |
|---|---|---|---|
| AI-generated code | Trade secret, with copyright only after substantial human rework | Open-source license contamination—licenses such as the GPL can pull obligations into the codebase | Substantial human editing by a developer, plus trade secret controls on access |
| Marketing copy | Copyright, only after human authorship | Competitors can legally reuse the same generated text, which enters the public domain as generated | Real human editing and arrangement before publication |
| Logo and UI design | Copyright through human-led design; trademark rights from use in commerce | The raw AI artwork can be copied and is weak as a brand asset | Human-led design for copyright; register and use brand elements in commerce for trademark |
| Fine-tuned models | Trade secret protection for model weights | Infringement claims from owners of the training data; reverse engineering | Use licensed or owned training data, plus confidentiality controls and contracts |
Two patterns run through the table. First, human intervention is the trigger for copyright: code, copy, and design each gain protection only when a person substantially reworks the raw output. Second, the assets that cannot be copyrighted—model weights in particular—shift toward trade secret protection, which makes confidentiality controls the operative safeguard rather than registration.
Patents vs. Trade Secrets: Strategic Choices for AI Innovation
For the underlying technology, the central decision is whether to patent an AI innovation or hold it as a trade secret. Shieh-Newton's analysis for Mintz explains why this choice is genuinely difficult for AI companies: the valuable intangible assets that provide competitive advantage can be hard to protect using IP filings such as patents, given an increasingly crowded market for AI innovations and a cluttered patent landscape.
The disclosure problem is the crux. A patent application must describe the invention, and in a worst-case scenario, Shieh-Newton warns, competitors may end up with a quasi-blueprint to practice the company's technology without the company holding enforceable exclusive rights. Fear of this scenario has contributed to some AI and technology companies investing less in patenting or avoiding patents altogether.
That does not make patents the wrong choice in every case. Shieh-Newton's guidance is that patent filing strategies should focus on patent quality over patent quantity, and that drawing the line between which AI innovations to patent and which to protect as trade secrets is one of the most important areas where a company benefits from expert, business-pragmatic IP advice. The decision factors are concrete: if a competitor can reverse-engineer the innovation from the product itself, secrecy alone offers little; if the innovation can be kept confidential through operations, trade secret protection may preserve advantage without public disclosure.
Trade secret protection, however, is not passive. Shieh-Newton emphasizes that for this protection to be effective, it must be intentional and operationalized through carefully designed business processes. Access controls, confidentiality agreements, and internal data-handling procedures are what convert a secret into an enforceable trade secret.
Navigating International IP Protection Challenges
Founders planning to sell outside their home market face an additional layer of complexity. IP commentary published by IPBA Connect on IP and AI startups highlights that each country has its own set of patent, copyright, trademark, and trade secret laws, which can differ in what is protectable, the duration of protection, and the procedures for obtaining and enforcing IP rights.
For AI specifically, the uncertainty is sharper. How different countries approach IP protection for AI innovations is far from uniform: the patentability of algorithms, the ownership of AI-generated works, and other legal gray areas all contribute to uncertainty across jurisdictions. A protection strategy that works in one market may not map onto another, and rights that depend on registration or use in a given country generally require action in that country.
The practical implication is timing. Because protection in each jurisdiction depends on that jurisdiction's rules, founders with international ambitions benefit from planning early—deciding which markets matter, what protection is available in each, and what filings or uses are required there—rather than retrofitting an IP strategy after expansion is underway.
Building a Comprehensive AI IP Portfolio
Assembled, the evidence points to a component-specific strategy rather than a single tool. Fine-tuned models sit at the center of the portfolio as trade secrets, protected through licensed or owned training data and confidentiality controls. Generated code needs substantial human editing to become protectable, and open-source license contamination must be policed before obligations attach. UI and logo assets gain copyright through human-led design, with trademark rights arising from use in commerce rather than from the artwork itself. Marketing copy requires real human editing and arrangement before publication to keep competitors from legally reusing it.
The patent-versus-trade-secret line should be drawn deliberately, innovation by innovation, with quality over quantity in any patent filings and carefully operationalized secrecy for everything held back. None of this constitutes legal advice for a specific situation, and the filing processes themselves—patent, trademark, or copyright—are separate undertakings beyond this framework's scope.
Developing Your AI IP Portfolio: Key Decisions
AI startup founders should proactively implement a tailored IP strategy that combines trade secrets, copyrights, and potentially patents and trademarks, aligning each protection type with specific AI assets and their unique vulnerabilities. The framework above is the working tool for that alignment: identify each asset, assign its protection, and act on the specific requirement—human editing, confidentiality controls, licensed data, or use in commerce—that makes the protection real.
The measurable indicator of a healthy strategy is a regular review of your IP asset inventory and protection measures, especially when developing new AI components or expanding into new markets. Each new model, codebase, or market entry changes the risk picture—new training data brings new infringement exposure, new jurisdictions bring different rules—and the review is the checkpoint that keeps the portfolio matched to the company as it grows.
