Artificial intelligence (AI) is transforming various industries, including the realm of intellectual property law. As AI systems become increasingly capable of generating content, recognizing brand elements, and automating commerce, new opportunities emerge for businesses. Copyright and trademark rights and potential infringement enforcement is one key area. Another is the increased reliance by businesses on AI for content creation, product design, and automation. AI is also impacting trademark registration searches for clearance. But, this rapid innovation also raises complex copyright and trademark challenges and many concerns. Understanding these issues is essential for businesses to protect their assets and leverage AI responsibly during this rapid evolution.
AI’s Role in Trademark Registration and Clearance
Traditional manual trademark clearance searches are time-consuming and can be prone to human error leading to overlooked conflicts. AI purports to transform this process by leveraging advanced technologies such as machine learning, natural language processing, and pattern recognition. Some of these tools claim to be able to analyze vast trademark databases across multiple jurisdictions and languages rapidly. Also, some AI models state the ability to search the Internet in real time, which can help identify conflicts. These AI tools should be able to detect not only exact matches but also subtle similarities in phonetics, visuals, and semantics that might be overlooked by human searchers, in theory. This capability can seemingly help reduce errors and ensure that new trademarks do not infringe on existing ones. While these models should not be used as a substitute for an experienced trademark attorney, they can help with efficiency and can speed up the process.
AI’s Role in Trademark Infringement and Enforcement
Several AI tools are also specifically designed to help businesses combat trademark infringement, each with unique functionalities. Tools like Visua.com and Huski.ai utilize machine learning algorithms to monitor digital spaces, including social media, e-commerce platforms, and domain registrations, for unauthorized use of trademarks. Various AI platforms such as Gemini or Grok 3 DeepSearch, etc. can be used to help search Internet records in real time to determine if any infringements may exist. These tools can be used to help identify infringements in real time, allowing businesses to respond swiftly in order to protect their brands.
Amazon’s implementation of AI for counterfeit detection serves as another practical example. The company uses AI to scan billions of listings daily, flagging products that may be counterfeit. In 2023, Amazon seized and disposed of over 7 million counterfeit products worldwide, demonstrating the scalability and effectiveness of AI in combating counterfeit goods. This approach includes automated tools within their Brand Registry, enhancing control over listings and reducing the time required to address infringements.
The future of AI in trademark protection is promising, with global spending on AI expected to reach USD 632 billion by 2028, driven by a 40.6% compound annual growth rate from 2023 (Norton Rose Fulbright on AI and IP). Proposals for new frameworks, such as “Digiwork” rights, suggest potential legal evolutions, but current emphasis on human involvement requires businesses to stay adaptable.
Challenges in Addressing AI-Driven Trademark Infringement
AI affects trademarks in multiple ways, from automated branding tools to AI-driven customer interactions. However, it also creates legal uncertainties in areas like counterfeit detection and brand misrepresentation. As AI systems become increasingly capable of generating content, recognizing brand elements, and automating commerce, new challenges emerge regarding trademark rights and potential infringement. For instance, AI-generated content, such as chatbots, virtual assistants, and automated marketing tools, may use brand names or logos without proper authorization. AI-powered e-commerce platforms using AI-driven search algorithms and recommendation engines can unintentionally favor counterfeit products or mislead customers about brand authenticity. may misclassify or recommend counterfeit goods, leading to consumer confusion and dilution of brand value. Consumers may mistakenly associate AI-generated products or services with established brands, impacting brand equity. AI can also create fake brand endorsements, deepfake videos, or counterfeit products, harming brand reputation and misleading consumers.
When AI systems infringe on trademark rights, determining liability becomes complex. Does responsibility fall on the AI developer, the user, or the company utilizing the AI system? AI continuously evolves, learning from vast amounts of data.
Are AI Generated Works Copyrightable?
If an AI generates artwork, text, or code, who holds the copyright? The programmer, the user, or no one? Determining ownership of AI-generated content is murky. In many jurisdictions, copyright protection applies only to human-created works, leaving businesses unsure of their rights over AI-produced material. Current laws, such as those in the US, require human authorship, as seen in the US Copyright Office’s January 2025 report, Part 2, which states, “Where that creativity is expressed through the use of AI systems, it continues to enjoy protection. Extending protection to material whose expressive elements are determined by a machine, however, would undermine rather than further the constitutional goals of copyright” (US Copyright Office Releases Part 2). This creates uncertainty for businesses using AI to generate content, as seen in the Kris Kashtanova case.
AI Legal Framework for Copyright Rights
The copyrightability of works created solely by AI is a contentious issue. The legal status of AI-generated images for commercial use hinges primarily on copyright law, which varies by jurisdiction but shares common themes globally.
- US Copyright Law: In the US, recent court rulings, such as the March 18, 2025 decision by the US Court of Appeals for the District of Columbia Circuit, have affirmed that AI-generated art without human input cannot be copyrighted (US appeals court rejects copyrights for AI-generated art lacking ‘human’ creator). This ruling involved Stephen Thaler’s AI system DABUS and reiterated that human authorship is a “bedrock requirement” for copyright protection. The US Copyright Office has also rejected applications for AI-generated images, like those from Midjourney, unless significant human creativity is involved (AI-generated art cannot receive copyrights, US court says).
- Human Input and Copyrightability: If a human contributes creatively, such as by providing detailed prompts, editing the output, or arranging AI-generated elements, they may claim copyright. For instance, the comic book “Zarya of the Dawn” saw its text and arrangement copyrighted, but not the AI-generated images themselves, as per the US Copyright Office’s guidance (New US copyright rules protect only AI art with ‘human authorship’). This suggests that the level of human involvement is critical, with prompts and modifications potentially qualifying as creative contributions.
- International Perspectives: The EU’s approach, outlined in the proposed AI Act, focuses on transparency regarding training data rather than ownership, requiring AI providers to disclose copyrighted materials used (Artificial intelligence and copyright: use of generative AI tools to develop new content). However, like the US, it does not explicitly address copyright for AI outputs, leaving it to national laws. The UK, notably, is one of the few jurisdictions allowing copyright for computer-generated works, but this is not standard globally (AI and Copyright Law: What We Know).
Can You Use AI-generated images/illustrations?
Whether AI-generated images can be used (presumably for commercial purposes) without copyright infringement concerns seems to depend on a few factors, including how the image was created, the tool used, and the terms of service of that tool. AI-generated images are typically created by models trained on vast datasets of existing images, often including copyrighted works, raising questions about fair use. Some argue this process involves “copying” or “derivative work,” while others see it as a transformative process that produces original output. OpenAI’s DALL-E states that illustrations it generates created using its OpenAI’s DALL·E model produces entirely new images based on text prompts. So, apparently, these images are not copied from existing works. Assuming this is accurate, then they are created based on learned patterns from a broad range of publicly available and licensed sources, but would not be direct reproductions of copyrighted material. AI-generated images and illustrations may be used on a blog without copyright issues if the tool you use grants rights to use the output and the content doesn’t mimic or include elements of a specific copyrighted work. Always review the tool’s terms of service (or ask it?), especially for commercial blogs and use for commercial purposes in general.
The following aspects should be considered:
- Tool Terms of Service: Many AI image generation tools (like Midjourney, DALL-E, or Stable Diffusion) specify who owns the output. Some grant you full commercial rights to use the images however you like, while others retain rights or impose restrictions (e.g., no commercial use without a paid license). The specific tool’s terms of use/service should always be reviewed. DALL-E tool and Gemini Advanced 2.0 Flash among others allow attribute and royalty free commercial use of images and illustrations created based on your inputs.
- Originality: If the AI produces an image that closely resembles an existing copyrighted work (e.g., a famous painting or photo), there appears to be a legitimate risk someone could claim infringement. Courts have started tackling this. Cases like Getty Images v. Stability AI (ongoing as of early 2025) have dealt with whether AI outputs can infringe on training data. Purely abstract or unique images would seem to be safer bets.
- Licensing: To be extra safe, use platforms that explicitly offer royalty-free or Creative Commons-licensed AI-generated images. Sites like Unsplash or Pixabay have started including AI-generated content under clear usage terms.
- Attribution: Even if not required, crediting the AI tool (e.g., “Image generated by Grok 3 from xAI”) can add transparency and may reduce risk.
What about Using AI Generated Text?
In order to use AI-generated text, like reports and articles, for commercial use is primarily a function of copyright law, and this is clearly still evolving per the key points specified below.
- US Copyright Law and AI-Generated Text: Recent guidance from the US Copyright Office, as outlined in their January 29, 2025, report on “Copyright and Artificial Intelligence, Part 2: Copyrightability,” states that outputs of generative AI can be protected by copyright only where a human author has determined sufficient expressive elements (US Copyright Office AI Report Part 2). This includes situations where a human-authored work is perceptible in the AI output or where a human makes creative arrangements or modifications, but not merely providing prompts. A report or article, for example, that is generated entirely based on a query with no further human input in the creation process seems unlikely to be copyrightable, as it lacks significant human authorship.
- Ownership and User Rights: If a given user commissions a report, for example, through interaction with AI, research suggests that user may likely have the right to use the output as he or she sees fit. The USF Libraries guide on AI tools indicates that if the terms of use do not address further uses of AI-generated material, the user is free to use it. However, this is a gray area, as AI-generated content’s ownership is still debated, with some arguing the user who provides the prompt may be considered the author, while others see it as public domain if no human creativity is involved.
- Comparison to AI-Generated Images: AI-generated images have been more widely litigated, such as the March 18, 2025, US appeals court ruling that AI-generated art without human input cannot be copyrighted (US appeals court rejects copyrights for AI-generated art lacking ‘human’ creator). AI-generated text follows similar principles, however. The case of the graphic novel “Zarya of the Dawn” showed that the selection and arrangement of AI-generated images could be copyrighted, but the images themselves could not (Baylor University Libraries AI Copyright FAQ). This suggests that if a user were to arrange or edit a research report creatively, the user might claim copyright, but reposting it as-is doesn’t involve such creativity.
- Training Data and Infringement Risks: A potential risk using an AI generated report as an example, while original in its analysis, it may summarize information from cited sources. If it is a summary and not a direct copy, fair use likely applies, but this is an area of ongoing controversy, with lawsuits like those against Stability AI highlighting training data issues (The Verge AI copyright concerns).
Practical Commercial Uses Come With Risks
Despite legal uncertainties, many businesses are already using AI-generated images commercially, attempting to adapt to the risks. Sectors like advertising, fashion, and publishing are adopting AI-generated images for cost-effective, rapid content creation. Examples include magazine covers (e.g., The Economist) and book illustrations, showing widespread use despite legal gray areas (Commercial image-generating AI raises thorny legal issues). Picsart and Midjourney both appear to allow commercial use under specific conditions. Picsart states that AI-generated images can be used commercially, but with disclaimers about copyright ownership and third-party rights (Are images generated through AI available for commercial use?). Midjourney, for instance, requires a paid subscription for commercial use, adding a cost barrier (Are AI-Generated Images Copyrighted). Platforms like Microsoft Copilot have commitments stating no restrictions on AI-generated materials (USF Libraries guide on AI Tools and Resources), suggesting a trend toward user freedom.
Training Data and Infringement Risks
Businesses can use AI generated content for commercial purposes, but they risk legal challenges if the images infringe on existing copyrights. Especially since AI models are trained on vast amounts of potentially copyrighted data. There seems to be a real risk that AI-generated images could resemble or copy existing copyrighted works, leading to disputes. This is particularly relevant given recent lawsuits over AI training data, such as with Stability AI, Midjourney and ROSS Intelligence. Numerous courts are still in the process of determining whether this constitutes fair use or copyright infringement, reflecting the ongoing and unresolved nature of these infringement disputes. The risk of liability remains on the business using AI platforms to generate content as these platforms often don’t guarantee copyright ownership or protection against third-party claims. Businesses are advised to review AI tool terms, ensure images don’t mimic copyrighted works, and consider legal counsel. Some platforms offer indemnification, but this is not universal, leaving businesses potentially liable for disputes.
Summary of Key AI Copyright & Trademark Legal Developments
Businesses and brand owners may gain insight in understanding the following recent developments concerning AI IP rights:
1. Perplexity AI’s Legal Challenges. In October 2024, Dow Jones and the New York Post filed a lawsuit against Perplexity AI, alleging copyright and trademark violations. The lawsuit claims that Perplexity fabricated and misattributed news content, leading to potential trademark dilution by confusing readers and harming publishers’ reputations. wired.com
2. Getty Images vs. Stability AI. In January 2023, Getty Images initiated legal proceedings against Stability AI in the English High Court, alleging unauthorized use of its images to train the Stable Diffusion model. Getty Images claims that Stability AI “scraped” millions of images from its websites without consent, infringing on both copyright and trademark rights. The case could potentially significantly impact AI training practices in the context of business IP rights.
3. Andersen et al. v. Stability AI Ltd. This was a class-action lawsuit by a group of artists against a few different AI Companies. In January 2023, artists Sarah Andersen, Kelly McKernan, and Karla Ortiz filed a class-action lawsuit against Stability AI, Midjourney, and DeviantArt. They alleged that these companies infringed upon the rights of millions of artists by using their works without consent to train AI tools. The lawsuit underscores the tension between AI development and the protection of individual creators’ intellectual property. theverge.com
4. News Publishers vs. AI Firm Cohere. In February 2025, over a dozen major U.S. news organizations, including Condé Nast and McClatchy, sued AI company Cohere. The lawsuit alleges that Cohere illegally used their publications to train its language model, resulting in copyright and trademark infringement. The publishers claim that Cohere’s practices harm their brands and seek damages for the unauthorized use of their content. Axios
5. Kris Kashtanova’s “Zarya of the Dawn”. In 2023, artist Kris Kashtanova used Midjourney’s AI to create images for the comic book “Zarya of the Dawn.” The US Copyright Office initially registered the work but later, on February 21, 2023, limited protection to the text and compilation, excluding AI-generated images, citing insufficient human authorship. As previously mentioned, businesses must ensure significant human creative input in AI-assisted works to qualify for copyright protection, a critical consideration for content creators. AI Comic Art Dispute
6. Thaler v. Perlmutter. As mentioned, this case involved the completely AI-generated work, “A Recent Entrance to Paradise”. Thaler, the copyright applicant, disclaimed any human involvement in the work’s creation. The U.S. District Court (D.C. Circuit) ruled it wasn’t copyrightable, stating that “the Copyright Act of 1976 requires all eligible work to be authored in the first instance by a human being.” Technology & Marketing Law Blog
7. Thomson Reuters Enterprise Centre GmbH v. ROSS Intelligence Inc. Filed in the U.S. District Court for the District of Delaware in May 2020, this suit alleges that copyrighted headnotes from Thomson’s Westlaw legal research database were used as training data for an AI legal research tool that was developed by ROSS Intelligence. (Recent Rulings in AI Copyright Lawsuits Shed Some Light, but Leave Many Questions)
8. Kadrey v. Meta Platforms (now consolidated with Chabon v. Meta Platforms). Meta’s LLaMa large language models are alleged to infringe upon derivative works because the models cannot function without the “expressive information” extracted from the plaintiff’s books. The U.S. District Court for the Northern District of California dismissed most of the plaintiffs’ claims, with only one plaintiff’s direct copyright infringement claim against Stability AI surviving. The court also rejected the plaintiffs’ theory that “every output of the model is an infringing derivative work” and that when third parties use the model, “every output …constitutes an act of vicarious infringement,” noting that the complaint offers no allegations that the output is infringing (i.e., that it is recasting, transforming, or adapting the plaintiffs’ books). The court granted leave to amend the complaint on most counts, and the plaintiffs have since filed an amended complaint in November 2023. (Recent Rulings in AI Copyright Lawsuits Shed Some Light, but Leave Many Questions)
These legal developments illustrate the evolving challenges at the intersection of AI development and IP rights. They also highlight the need for updated legal frameworks to address potential infringements in this rapidly changing landscape. One trend seems to be that claims that the large language models themselves and that the generated output of such models are infringing works is not a valid claim. Unless the output contains portions of the copyrighted content or portions that are substantially similar to the copyrighted content.
Conclusion
As AI continues to reshape commerce and branding, businesses must proactively adapt to protect their copyright and trademark rights. AI presents a dual-edged sword for businesses, offering tools for enhanced protection while introducing legal uncertainties. The best practices dictate i) utilizing AI-powered trademark monitoring tools to detect and prevent potential infringements and ii) establishing clear policies for AI-generated content to avoid unintentional trademark misuse. As the legal framework evolves, staying informed and proactive will be key to balancing innovation with IP protection, ensuring businesses thrive in an AI-driven world. By understanding case studies like the Kris Kashtanova issue, and implementing best practices, businesses can navigate the AI landscape effectively.
Looking ahead, AI is poised to introduce further innovations in trademark law, such as dynamic trademarks that adapt to market changes, offering brands greater flexibility. As AI evolves, it may become reliable in interpreting search results and providing reliable opinions on trademark registration. Additionally, integrated global systems could streamline trademark registration and enforcement across jurisdictions, making it easier for businesses to protect their IP rights on a global scale. Embracing these advancements while addressing the associated legal and ethical concerns will be key to unlocking the full potential of AI in business IP copyright and trademark rights management.