Artificial intelligence has quickly become a practical marketing tool for small businesses. A local retailer can use AI to draft product descriptions, a professional services firm can ask a chatbot for blog topics, a restaurant can generate social media captions, and an online store can experiment with AI-generated images for advertisements. These tools can save time, lower creative costs, and help small businesses publish more consistently. They also raise a deceptively simple legal question: who owns the content once AI helps create it?
The answer is not always what business owners expect. Under current United States copyright principles, ownership begins with authorship, and copyright authorship still depends on human creativity. The U.S. Copyright Office has explained that copyright protects human-authored expression, not material produced by a machine without sufficient human creative contribution. AI may be used as a tool in the creative process, but copyright protection depends on whether a human being made the expressive choices reflected in the final work. In other words, the fact that a person typed a prompt does not automatically mean that person owns copyright in everything the AI system produced. ¹
For small businesses, the important question is not simply whether a person paid for the AI tool, opened the account, or entered the prompt. The better question is what the human actually contributed to the final marketing asset. A business owner who asks AI to “write a Facebook post about our spring sale” and publishes the result unchanged may have little or no copyright ownership in the AI-generated wording itself. A business owner who uses AI to brainstorm, then rewrites the copy, chooses the structure, adds original examples, adjusts the tone, and creates the final message may own the human-authored parts of the finished work. The more the final content reflects human judgment, selection, arrangement, editing, and original expression, the stronger the copyright position becomes.
This distinction matters because “owning” marketing content can mean several different things. A small business may have permission under an AI platform’s terms of service to use the output commercially. It may own its own human edits, campaign strategy, brand voice, photography, logo, layout, or design choices. It may also have a written contract with a marketing agency assigning rights in the deliverables. But those facts do not automatically mean the business owns copyright in raw AI-generated material. Copyright ownership, platform permission, contract rights, trademark rights, and practical control over a file are separate concepts. Confusing them can create problems when a business later wants to register a work, stop a competitor from copying it, sell a brand package, license content, or prove that it owns the assets used in a campaign.
The Copyright Office’s registration guidance is especially important for small businesses that want to register marketing materials containing AI-generated content. The Office has explained that when AI determines the traditional expressive elements of a work, the generated material itself is not protected by copyright. The Office has also stated that applicants should disclose AI-generated content in registration applications and identify the human author’s contribution. A business may claim copyright in human-authored text, creative selection and arrangement, or substantial human modifications, but AI-generated material that is more than minimal generally should not be claimed as human-authored expression. ²
The courts have reinforced the same human-authorship principle. In Thaler v. Perlmutter, the United States Court of Appeals for the District of Columbia Circuit affirmed the Copyright Office’s refusal to register an artwork where the AI system was identified as the author. The court treated human authorship as a requirement of copyright protection under the Copyright Act.³ This does not mean every AI-assisted marketing asset is unprotectable. It means a business should not assume that AI output, standing alone, carries the same copyright protection as a human-written advertisement, a human-designed brochure, or a human-shot photograph.
The practical rule is that AI should be treated as a tool, not as the legal author. A camera can help create a copyrightable photograph because the photographer may choose the subject, lighting, angle, timing, framing, and other creative elements. Design software can help create a copyrightable advertisement because the designer may make original choices about layout, color, typography, and composition. AI can also be part of a human creative process. But where AI independently supplies the expressive wording, image, melody, or layout, and the human merely accepts the output, the copyright claim becomes weaker.
This is especially important for marketing copy. Many businesses now use AI to generate blog posts, landing pages, emails, ad headlines, product descriptions, and social media captions. A lightly prompted AI output may be useful as a draft, but the safer copyright practice is to treat it as raw material. A business should have a human review the draft, revise the substance, adjust the voice, add specific knowledge, remove generic phrasing, verify factual claims, and shape the final expression. A published blog post that reflects human analysis, examples, organization, and editing is much more defensible than a pasted AI response. The human work does not need to be dramatic, but it should be real, original, and visible in the final content.
Images create a similar issue. If a business uses AI to generate a background image for an advertisement and publishes it unchanged, the business may have permission to use the image under the platform’s terms, but that does not necessarily mean it owns copyright in the image. If a designer uses AI to create rough concepts, then combines them with original photography, custom typography, brand colors, hand-created graphics, and a carefully designed layout, the human-authored portions and the creative arrangement may be protectable. The more the final image reflects identifiable human choices, the better the business’s position.
Logos require special caution. A logo is often one of the most valuable marketing assets a small business owns. Copyright may protect some logo artwork if it contains enough original expression, while trademark law may protect the logo as a brand identifier. But if a logo is generated primarily by AI, the business may face uncertainty about copyright ownership in the artwork. The generated logo could also resemble existing designs, contain unoriginal elements, or be difficult to register as a copyright claim. A business that wants a durable logo should involve a human designer, document the design process, conduct clearance review, and make sure contracts assign rights in the final design.
The ownership question becomes more complicated when employees use AI. If an employee creates marketing content within the scope of employment, the employer may own the copyright under the work-made-for-hire doctrine. The Copyright Office explains that a work made for hire can exist when an employee creates a work as part of the employee’s regular duties, and in that situation the employer is generally considered the author and copyright owner.⁴ In the AI context, however, the employer can only own copyright in copyrightable human-authored material. If the employee merely causes AI to generate content without sufficient human authorship, there may be little copyright for the employer to own. If the employee uses AI as a drafting or design aid and then contributes original expression, the employer’s ownership position is much stronger.
Independent contractors and marketing agencies require even more care. Many small businesses assume that paying a freelancer means owning everything the freelancer creates. That assumption can be wrong. Work-made-for-hire rules for commissioned works are narrower than many business owners realize, and written agreements matter. A commissioned work generally needs to meet specific legal requirements before it qualifies as a work made for hire, including an appropriate written agreement between the parties.⁴ Without the right contract language, a small business may end up with only an implied or limited license rather than full ownership.
For that reason, a business hiring a designer, copywriter, social media manager, videographer, or advertising agency should use a written contract that addresses AI directly. The contract should not merely say that the business owns “the content.” It should define the deliverables, require disclosure of AI-generated portions, assign all copyrightable human-authored contributions, give the business broad rights to use any noncopyrightable or AI-generated portions, require the contractor to follow platform terms, and require the contractor to avoid knowingly using infringing inputs. If the contractor is expected to provide original human authorship, the agreement should say so. If the contractor may use AI only for brainstorming or rough drafting, that should also be clear.
Assignments are also important. Under federal copyright law, a transfer of copyright ownership generally must be in writing and signed by the owner of the rights being transferred or by that person’s authorized agent.⁵ This rule matters when a small business wants to buy a website design, a brand guide, a video script, a photo package, or a set of advertisements. Even if the finished campaign contains AI-assisted material, there may be human-authored elements that need to be assigned. A written assignment helps the business avoid later disputes over who can reuse, modify, license, or enforce the work.
Platform terms are another layer, but they should not be mistaken for copyright law. Many AI tools tell users they may use outputs commercially or that the user owns the output as between the user and the platform. That type of term can be valuable because it may mean the platform is not claiming ownership against the user. It may also permit commercial use in advertisements, websites, emails, packaging, and social media posts. But platform terms cannot turn noncopyrightable machine-generated expression into copyrightable human authorship. They also may not protect the business if an output resembles someone else’s work, includes a third party’s trademark, violates publicity rights, or was created using restricted inputs. A business should read the terms of the specific tool it uses and should not assume all AI tools provide the same rights.
Small businesses should also consider similarity risk. AI systems can generate outputs that are generic, repetitive, or similar to outputs received by other users. A marketing slogan, product description, or image generated from a common prompt may not be unique. Even if the business has permission to use the output, it may not be able to stop another business from using similar content. This is a practical problem as much as a legal one. Marketing works best when it is distinctive. A business that relies too heavily on unedited AI output may end up with content that is not only difficult to protect, but also forgettable.
There is also the risk of accidental infringement. A business should not ask AI to write in the exact style of a living author, generate an image just like a specific artist’s protected work, copy a competitor’s advertisement, imitate a celebrity endorsement, or rewrite someone else’s copyrighted blog post while keeping the same expressive structure. Even when the final output is not identical, the prompt itself may encourage the system to produce material that is too close to protected expression. The safer practice is to use AI for general brainstorming, drafting, and editing while avoiding prompts that request imitation of specific protected works or identifiable living creators.
For marketing teams, documentation is one of the best risk-management habits. A business should keep records showing how a final asset was created. This does not require preserving every draft forever, but it is wise to save meaningful evidence of human contribution. Drafts, edits, notes, design files, prompt logs, revision history, and comments can help show that a human selected, arranged, modified, and finalized the work. If the business later registers the content, sells the company, licenses the campaign, or responds to a dispute, those records can be useful.
Registration strategy should be thoughtful. Not every social media post or email needs a copyright registration. But a business may want to register major assets such as a website, flagship guide, video campaign, catalog, original photography collection, training course, or substantial brand publication. If the work includes AI-generated content, the registration should be prepared carefully. The application should identify the human author or authors, describe the human-authored material, and exclude AI-generated material where required. The goal is not to hide AI use, but to define the copyright claim accurately.
The same principle applies to website content. A modern small business website may include AI-assisted copy, stock photography, original photography, icons, template elements, human-designed layouts, embedded videos, and customer testimonials. The business may own some elements, license others, and have no copyright in still others. A website is therefore not a single ownership question. It is a bundle of rights. A careful business will know which parts were created by employees, which were created by contractors, which were licensed from third parties, which were generated by AI, and which were substantially revised by humans.
Social media marketing adds another wrinkle because speed can undermine review. A business may be tempted to generate and post content quickly, especially for daily captions or trend-based posts. That is understandable, but speed should not eliminate human oversight. Someone should verify factual claims, confirm that promotional statements are accurate, check whether images contain confusing trademarks or recognizable people, and make sure the post reflects the brand’s actual voice. Copyright ownership is only one issue. Advertising claims, privacy, publicity rights, platform rules, and reputation all matter.
Email marketing deserves similar attention. AI can help draft subject lines, promotional sequences, and customer newsletters, but a business should avoid sending generic AI-generated claims without review. If the email describes a sale, warranty, service promise, price, testimonial, or comparison to competitors, a human should confirm accuracy. From a copyright perspective, human revision can improve ownership. From a business perspective, human review can prevent misleading statements and preserve customer trust.
For paid advertising, the stakes are higher because the content is designed to generate revenue and may be seen widely. A business using AI-generated ad copy or images should ask whether it can prove the right to use the material, whether the material is distinctive, whether it accidentally resembles a competitor’s branding, and whether the final ad includes enough human authorship to support ownership. A business should be especially careful before using AI-generated images of people, products, buildings, artwork, or recognizable brand-like symbols. Copyright is not the only concern. Trademark, right of publicity, privacy, and false endorsement issues can also arise.
Small businesses should not respond to these risks by avoiding AI entirely. The better response is to use AI deliberately. AI is excellent for brainstorming blog topics, generating outlines, suggesting alternative headlines, summarizing research, creating rough first drafts, identifying gaps in copy, and proposing variations for human review. AI is less reliable as the sole source of final creative expression. The strongest approach is to let AI accelerate the process while allowing humans to control the final voice, substance, arrangement, and expression.
A good internal AI policy can be simple. It should tell employees which tools are approved, what information may not be entered into AI systems, when AI use must be disclosed internally, who reviews AI-assisted content before publication, and how the company documents human edits. It should also address whether employees may use AI-generated images, whether legal review is needed for major campaigns, and how contractors must report AI use. A policy does not need to be long to be useful. It needs to match the business’s actual workflow.
Confidentiality should be part of that policy. Marketing teams often work with unreleased products, customer lists, pricing plans, advertising strategies, internal data, and brand concepts. Entering sensitive information into public AI tools may create privacy, confidentiality, or trade secret concerns, depending on the tool and the circumstances. A business should not put confidential customer information, unreleased financial data, private employee information, or proprietary plans into an AI system unless it understands the tool’s data-use terms and has appropriate safeguards in place.
The human-authorship rule also has a strategic upside. It encourages businesses to make their marketing more original. A company that merely publishes AI output may receive something fast but generic. A company that uses AI to support human creativity can develop a stronger voice and more defensible assets. The legal and branding incentives point in the same direction. Human judgment makes the work more protectable, and human judgment also makes the work better.
The most practical answer to “who owns AI-generated marketing content?” is layered. If the content is generated entirely by AI with no meaningful human creative contribution, the business may have permission to use it, but it may not own copyright in the output. If a human employee uses AI as a tool and creates original expression within the scope of employment, the business may own the copyrightable human-authored portions as a work made for hire. If a freelancer or agency creates the content, ownership depends heavily on the written contract and assignment language. If the final work combines AI-generated material with human-authored text, design, selection, arrangement, or modification, copyright may protect the human-authored portions while excluding the AI-generated portions.
For small businesses, the safest operating rule is straightforward: do not rely on AI alone for important marketing assets. Use AI to assist, but make sure a human meaningfully edits and finalizes the work. Use written contracts with agencies and freelancers. Read platform terms. Keep records of human contributions. Be accurate in copyright registrations. Avoid prompts that ask AI to imitate protected works or living creators. Treat major brand assets, logos, websites, ad campaigns, and videos as legal assets worth managing carefully.
AI can be a powerful marketing assistant, but it is not a substitute for ownership planning. The businesses that benefit most from AI will be the ones that combine efficiency with discipline. They will use AI to move faster, but they will preserve human creativity, document the process, and secure rights through contracts. In the current legal landscape, that is the best way for small businesses to build marketing content they can use confidently, protect where possible, and rely on as part of their long-term brand value.
Contact Tishkoff
Tishkoff PLC specializes in business law and litigation. For inquiries, contact us at www.tish.law/contact/. & check out Tishkoff PLC’s Website (www.Tish.Law/), eBooks (www.Tish.Law/e-books), Blogs (www.Tish.Law/blog) and References (www.Tish.Law/resources).
Footnoted Sources
1- U.S. Copyright Office, Copyright and Artificial Intelligence, Part 2: Copyrightability, January 29, 2025. https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdf?utm_source=chatgpt.com
2- U.S. Copyright Office, Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Federal Register 16190, March 16, 2023. https://www.federalregister.gov/documents/2023/03/16/2023-05321/copyright-registration-guidance-works-containing-material-generated-by-artificial-intelligence
3- Thaler v. Perlmutter, No. 23-5233, United States Court of Appeals for the District of Columbia Circuit, March 18, 2025. https://media.cadc.uscourts.gov/opinions/docs/2025/03/23-5233.pdf https://www.supremecourt.gov/search.aspx?filename=%2Fdocket%2Fdocketfiles%2Fhtml%2Fpublic%2F25-449.html
4- U.S. Copyright Office, Circular 30: Works Made for Hire, revised August 2024. https://copyright.gov/circs/circ30.pdf
5- 17 U.S.C. § 204, Execution of Transfers of Copyright Ownership, United States Code. https://uscode.house.gov/view.xhtml?edition=prelim&num=0&req=granuleid%3AUSC-prelim-title17-section204
Top of Form
Bottom of Form