By Brett Trout
Lawyers across the country are using artificial intelligence (AI) to generate documents, images, videos, audio, software code, simulations, and 3D designs. This type of “generative AI” is no longer a novelty; it has become a core business tool reshaping industries from marketing and manufacturing to medicine and law. For lawyers, it presents both opportunity and risk. Nowhere is that tension more pronounced than in intellectual property (IP) law, where doctrines built for human creativity are being tested by machines capable of generating content at extraordinary speed. For Iowa practitioners, whether in their own practices or advising businesses, creators, and innovators, mastering the intersection of IP and generative AI is quickly becoming essential. This article explores the key legal principles, emerging trends, and practical considerations every lawyer should understand in navigating this evolving landscape.
At its core, U.S. intellectual property law is premised on human authorship and invention and generative AI disrupts that foundation. Modern AI systems are trained on vast datasets that often include copyrighted works and can produce outputs that recombine and sometimes resemble existing material. This raises fundamental questions about who owns AI-generated content, whether training on copyrighted material constitutes infringement, and who bears liability when AI outputs violate IP rights. Courts, regulators, and practitioners are grappling with these issues in real time, and existing IP frameworks were simply not designed with non-human creators in mind, leaving uncertainty about how rights can be protected or enforced.
Copyright law
Copyright law has become the front line of disputes involving generative AI. The U.S. Copyright Office has made clear that copyright protection requires human authorship, meaning works generated entirely by AI without meaningful human input are not eligible for protection. This creates an important distinction for practitioners: purely AI-generated outputs are generally not copyrightable, while works that are edited, curated, or otherwise shaped by human creativity may qualify for protection to the extent of that human contribution. Even minimal creative input, such as selection, arrangement, or editing, may satisfy the threshold for copyright protection. At the same time, clients often assume that crafting detailed prompts confers ownership rights, but current guidance indicates that prompting alone, even when iterative, typically does not constitute sufficient authorship. This has important implications for contracts, licensing, and ownership disputes, and lawyers should caution clients against assuming rights attach to AI prompts or outputs.
Generative AI also creates significant infringement risks because it can produce outputs that are substantially similar to protected works. Recent disputes have highlighted this risk, particularly where AI-generated images mimic copyrighted content. In many cases, derivative works such as summaries, remakes, prequels, and sequels may infringe underlying copyrights even if no exact text or images are copied. As a result, traditional infringement analysis, including substantial similarity, access, and fair use, remains highly relevant but must now be applied in new contexts involving AI. When AI outputs resemble existing works or incorporate recognizable characters or settings, obtaining a legal opinion before use is a prudent step.
The issue of training data represents one of the most consequential and unsettled areas in AI-related IP law. A central question is whether training AI models on copyrighted material constitutes infringement. Developers often argue that training qualifies as fair use because it transforms underlying data into new outputs, while critics contend that large-scale ingestion of copyrighted works without permission is unlawful. This debate remains unresolved but is at the center of ongoing litigation and legislative efforts. At least one court has drawn a distinction between training on purchased materials, which may qualify as fair use, and training on non-purchased materials, which may not. However, outcomes remain inconsistent, and the rapidly evolving nature of the law means that what is accurate today may quickly become outdated. Practitioners must stay current with Copyright Office guidance and federal court decisions when advising clients.
Trade secrets
Trade secret law introduces another layer of complexity, particularly regarding confidentiality. While it is possible to obtain trade secret protection for material created with AI, a more immediate risk arises when sensitive information is shared with AI platforms that may train on and disseminate that data. Recent federal decisions have made clear that voluntarily disclosing proprietary information to an AI platform can defeat trade secret protection, and communications involving AI tools may lose attorney-client privilege and become discoverable in litigation if confidentiality is not preserved. For lawyers, this underscores a critical point: generative AI should never be assumed to be a confidential environment. Using AI tools that train on user inputs jeopardize both trade secrets and privilege. Law firms and clients must implement strict written policies governing AI use, particularly when handling sensitive information, and these policies must be regularly updated to reflect changes in the law and technology.
Patents
The U.S. Patent and Trademark Office has rejected the idea that AI systems can be inventors, maintaining the requirement of human inventorship. However, AI-assisted inventions raise complex questions about the degree of human contribution necessary and who qualifies as the inventor when AI plays a significant role. These issues are likely to generate increased litigation in the coming years, and practitioners should advise clients developing AI-assisted technologies to maintain detailed records distinguishing human and AI contributions.
Trademarks
In the trademark context, it is possible to obtain protection for words or logos generated by AI, but the risk of infringing existing trademarks remains. Additionally, generative AI raises concerns in advertising related to rights of publicity and privacy, as it can create realistic depictions of individuals, including their likeness, voice, and mannerisms. This has led to issues involving false endorsements and deepfakes. While several states have enacted legislation addressing these concerns, federal regulators are exploring broader, more uniform approaches.
Liability
Liability for AI-related IP violations is another evolving area. Although much of the responsibility may fall on AI developers, users, including lawyers and their clients, can also face significant exposure, particularly when AI-generated content is used in commercial contexts such as marketing, software development, or publication. While infringement claims may still rely on traditional doctrines such as substantial similarity, derivative works, and likelihood of confusion, they will be applied in novel ways. As a result, lawyers should incorporate contractual protections such as warranties, usage restrictions, and indemnification provisions, and should carefully review AI-generated outputs before use or distribution.
Best practices
Lawyers advising clients on protecting AI-generated intellectual property should focus first on maximizing the role of human contribution. Because current U.S. law ties copyright and patent rights to human authorship and inventorship, attorneys should encourage clients to meaningfully edit, curate, and document their involvement in any AI-generated output. Maintaining clear records that distinguish between human and machine contributions can be critical in establishing ownership and enforceability. Even when generative AI output is not protectable, post-processing even a small amount of human input into the AI output makes it difficult for competitors to identify and extract the non-protectable portions of the output. Where appropriate, lawyers should also consider alternative protections such as trade secrets, contractual restrictions, and licensing frameworks, particularly when the output itself may not qualify for traditional IP protection.
Equally important is implementing strong contractual and policy-based safeguards. Lawyers should ensure that agreements with employees, contractors, and vendors clearly address use and ownership of AI-assisted work product, including assignment provisions and warranties regarding the use of AI tools. Clients should also adopt written internal AI use policies that restrict the input of confidential or proprietary information into public systems and define acceptable use of generative tools. In commercial transactions, attorneys should negotiate representations, warranties, and indemnification provisions that allocate risk related to AI-generated content, particularly where third-party rights may be implicated.
To avoid infringing third-party intellectual property, lawyers must advise clients to treat AI outputs with the same scrutiny as any third-party content. This includes reviewing outputs for substantial similarity to existing works, avoiding prompts designed to replicate specific copyrighted materials or identifiable brands, and conducting clearance searches when using AI-generated content in commerce. Clients must also be cautious about the provenance of AI tools themselves, including how they are trained and whether their terms of use provide adequate protections. Ultimately, effective risk management requires combining traditional IP diligence with a heightened awareness of how generative AI can introduce hidden infringement risks.
The intersection of intellectual property and generative AI remains in its infancy, with courts issuing early decisions and regulators and legislators working to define the legal framework. What is clear is that AI will not fit neatly into existing doctrines, and the law will continue to evolve, whether incrementally or through significant reform. New cases are emerging rapidly, shaping the contours of AI-related IP law in real time. For Iowa lawyers, generative AI is not a distant issue but a present reality. Whether advising startups, agricultural innovators, healthcare providers, or creative professionals, attorneys must understand how AI intersects with intellectual property rights. Human input remains the cornerstone of copyright and patent protection, but AI introduces new risks related to infringement and ownership. Careless use of AI can compromise confidentiality and privilege, and the legal landscape continues to shift with no settled answers. Mastery of these issues is no longer optional; it is essential to competent and forward-looking legal practice in the age of generative AI.
About the author:
Brett J. Trout is an Iowa patent lawyer with the BrownWinick Law Firm.
