What AI “Thinks About” Its Own Impact on IP: Three Questions with Microsoft Copilot

Microsoft Copilot’s visual representation of itself

The impact of artificial intelligence (AI) on intellectual property (IP) has been one of the most discussed topics within the global IP community in recent years. Here, we give AI the opportunity to weigh in on this evolving debate.

In this new interview, Microsoft Copilot responds to three questions about the intersection of AI and IP, offering its perspective on how the technology has impacted IP so far, what changes to expect within the next five years, and how AI may eventually reshape IP.

What has been the impact of generative AI on IP since ChatGPT became publicly available?
Since the public release of ChatGPT and other generative AI tools, the impact on IP has been profound and multifaceted. Here are the key developments and challenges:

Copyright Infringement and Training Data: Generative AI models are trained on vast datasets, often scraped from the Internet. This has raised concerns about unlicensed use of copyrighted material in training data and about the lack of transparency about what data was used, making it difficult to assess potential IP violations.

Several lawsuits have been filed by artists, authors, and media companies alleging that their copyrighted works were used without permission to train AI models.

Ownership of AI-Generated Content: A major legal gray area pertains to who owns the output of generative AI. In many jurisdictions, AI-generated works are not eligible for copyright unless there is significant human input. This creates uncertainty for businesses and creators using AI to produce content, as they may not be able to claim exclusive rights.

Risk to Proprietary Information: Organizations using generative AI tools risk inadvertently exposing their own IP. Prompts or data entered into AI systems could be retained or used to further train models. This has led to concerns about data leakage and loss of trade secrets, especially in enterprise settings.

Creation of New IP: On the flip side, generative AI has enabled the rapid creation of new content and innovations, from marketing copy to software code and product designs. This has accelerated innovation cycles and raised questions about how to protect AI-assisted inventions under existing IP frameworks.

Legal and Regulatory Responses: Courts and lawmakers are beginning to respond. Ongoing litigation is shaping how copyright law applies to AI, and in some jurisdictions are exploring new legal frameworks to address AI-generated content and training data rights.

How do you see the relationship between AI and IP evolving over the next five years?
Over the next five years, the relationship between AI and IP is expected to evolve significantly, driven by legal, technological, and societal shifts. Here’s a forward-looking view based on current trends and expert analysis:

The Solidification of Legal Frameworks: New legislation and case law will likely define clearer rules around AI-generated content. Courts are already grappling with questions of authorship and ownership, and landmark decisions are expected to set precedents. Jurisdictions such as the EU and UK are leading efforts to codify rules for AI-generated works, often assigning authorship to the person who configures or directs the AI.

AI as a Tool for IP Enforcement: AI will increasingly be used to detect and prevent IP infringement, such as identifying unauthorized use of copyrighted images, music, or text online. Tools powered by AI will help automate IP portfolio management, monitor global markets for counterfeits, and streamline litigation.

Licensing and Training Data Transparency: There will be growing pressure on AI developers to disclose training data sources and obtain proper licenses. We can expect to see standardized licensing models for using copyrighted content in AI training, possibly similar to how music licensing works today.

Rise of AI-Created IP and New Business Models: As AI becomes more capable, it will generate patentable inventions, creative works, and trademarks—raising questions about how to protect and monetize these assets. Businesses may develop hybrid human-AI co-creation models, where human input is used strategically to secure IP rights.

Global Divergence in IP Policy: Different governments may adopt divergent approaches to AI and IP, leading to jurisdictional complexities. For example, the U.S. may continue to deny copyright to purely AI-generated works, while other governments might allow limited protection under specific conditions.

Ethical and Economic Implications: The IP system will need to balance incentives for human creators with the growing capabilities of AI. There may be calls for new categories of rights or protections tailored to AI-generated content, especially in creative industries.

Finally, in your view, will AI bring about the end of IP as we know it?
AI is unlikely to bring about the end of IP, but it is certainly transforming it in ways that may lead to a fundamental redefinition of what IP means and how it functions. Here’s a breakdown of how this transformation might unfold:

Evolution, Not Extinction: Rather than eliminating IP, AI is forcing it to evolve. Traditional IP laws were designed for human creators and inventors. AI challenges these assumptions by generating content, inventions, and designs autonomously or semi-autonomously. We’re likely to see new categories or adaptations of IP law rather than its disappearance.

Shift in Legal and Ethical Boundaries: Regarding copyright, we may see a shift toward protecting the prompt engineer or human collaborator rather than the AI itself. On the other hand, patents could require new standards for what constitutes an “inventive step” when AI is involved. And trademarks might be impacted by AI-generated branding or deepfakes.

This could lead to hybrid legal frameworks that blend human and machine contributions.

Redefining Creativity and Ownership: AI blurs the line between originality and imitation on one hand, and human authorship and machine generation on the other. This raises philosophical and legal questions about what it means to create, and whether ownership should be tied to effort, intent, or innovation.

Global Fragmentation: Different countries are taking divergent approaches. The U.S. leans toward denying copyright to AI-generated works. The EU and UK are exploring more nuanced or flexible models. And China is investing heavily in AI IP protections to support innovation.

This could lead to fragmented global IP regimes, complicating international enforcement.

New Opportunities and Risks: Opportunities include faster innovation, democratized creativity, and new business models, while risks include IP theft, data misuse, and erosion of human creative industries. The challenge will be to balance innovation with protection, ensuring that creators, inventors, and the public all benefit.

Final Thought?
AI won’t end IP—it will reshape it. The next five years will likely see a reimagining of IP rights to accommodate a world where machines are not just tools, but collaborators.

Although every effort has been made to verify the accuracy of this article, readers are urged to check independently on matters of specific concern or interest.

© 2025 International Trademark Association