Navigating Intellectual Property Challenges in Protecting AI Assets: Corporate Governance Imperatives

Answer first: The global valuation of intangible assets surged to $57.3 trillion in 2023, with AI increasingly integral to innovation across sectors. However, protecting AI-related intellectual property presents complex challenges due to unsettled patent and copyright laws. This article outlines key considerations for corporate boa…

AI & IP context for IP teams

In 2023, the worldwide value of intangible assets—including contractual rights and intellectual property such as patents, copyrights, and proprietary technology—reached an unprecedented USD 57.3 trillion. This figure not only surpassed pre-pandemic levels but also marked an 8% increase from 2022, with the United States contributing USD 836 billion to this year-over-year growth. Major industries, ranging from pharmaceuticals to telecommunications, now boast intangible asset valuations exceeding the trillion-dollar mark.

Intellectual property-intensive sectors are increasingly leveraging artificial intelligence (AI) to accelerate research and development of novel products and services. For instance, leading pharmaceutical companies integrate AI into laboratory workflows to gain competitive advantages. As AI technology advances, it is poised to significantly influence corporate valuations of intangible assets, potentially impacting trillions of dollars in value. Given this high stake, it is critical for corporate board members and their advisors to stay informed on the latest legal developments concerning AI-related asset protection and risk mitigation strategies.

Key takeaways for protecting AI-related intellectual property

• Confirm how the development affects ai & ip ownership, enforcement, licensing, or portfolio records.
• Separate confirmed facts from legal interpretation before advising business teams.
• Map deadlines, affected assets, contracts, and evidence files to the responsible internal owner.
• Use the issue as a prompt for monitoring, filing strategy, dispute preparation, or member education.

Practical analysis

Despite familiarity with traditional IP protections such as patents and copyrights, companies face uncertainties when applying these frameworks to AI models and their outputs. Patents, which provide public documentation of inventions and define protection scope, remain a key mechanism for safeguarding AI innovations, as evidenced by the increasing issuance of AI-related patents. However, challenges persist. In the United States, questions remain about whether AI software, models, or platforms qualify as patent-eligible subject matter. Even if eligibility is established, the complex and evolving nature of AI models complicates the description and claim drafting process. Moreover, the opaque "black box" characteristic of AI systems hampers infringement analysis against competitors’ models.

Patent protection for AI-generated outputs or applications may circumvent some hurdles faced by AI model patents but introduces distinct issues. Current laws in the U.S. and Europe prohibit listing AI as an inventor on patents, rendering applications without human inventive contribution nonviable unless legal reforms occur. The treatment of inventions involving both human and AI contributions remains uncertain. Additionally, the standard of non-obviousness—central to patentability—may require reevaluation, as AI’s capacity to analyze vast knowledge bases challenges traditional human-centric assessments. These ambiguities complicate companies’ ability to accurately value their own and competitors’ AI-related IP portfolios.

Copyright law also presents complexities for AI. U.S. courts and the Copyright Office have determined that AI cannot be recognized as an author, rejecting claims that human prompts alone constitute authorship. Consequently, elements of creative works attributable solely to AI lack copyright protection. Furthermore, training generative AI models necessitates access to extensive datasets, some containing copyrighted materials such as images and literary works. Numerous ongoing litigations worldwide contest whether such training infringes copyrights or qualifies as fair use, with resolutions expected to take years.

In light of these uncertainties, companies increasingly consider trade secrets and contractual agreements to protect AI assets. Trade secrets safeguard confidential information that holds commercial value by virtue of secrecy and reasonable protective measures. While avoiding patent-related obstacles, trade secrets do not prevent competitors from independently developing or reverse-engineering similar AI models. Publicly accessible AI models further diminish trade secret efficacy. Additionally, maintaining secrecy can obscure the value of AI assets to external stakeholders and partners.

Contractual provisions enable companies to delineate ownership, usage rights, and indemnification responsibilities among collaborators concerning AI models and outputs. However, contracts only govern relationships between parties and offer no protection against third-party infringement. Therefore, companies must carefully craft agreements to address these issues while recognizing their limitations.

Given the absence of a universal solution, companies must tailor IP strategies for AI assets based on product specifics, market context, and end-user considerations. Board members should ensure that corporate leadership evaluates the advantages and drawbacks of various protection mechanisms and implements comprehensive protocols to safeguard and monetize AI innovations.

A foundational step involves establishing a robust AI policy that articulates IP considerations and preferred protection approaches. Such a policy should encompass the full AI ecosystem, including hardware, algorithms, models, and outputs. While new AI hardware and applications may qualify for patent protection, AI algorithms and models might be better protected as trade secrets. The policy should emphasize procedural safeguards to maintain protection regardless of the chosen strategy.

To facilitate patenting, companies should meticulously document human involvement throughout AI development stages—ranging from conceptualization and hardware design to algorithm creation, data selection, prompt generation, output interpretation, and utilization. This documentation aids in identifying inventors and securing appropriate rights assignments.

Maintaining AI as a trade secret requires implementing reasonable confidentiality measures, restricting access to authorized personnel, and securing agreements obligating employees and third parties to uphold secrecy. The policy should also address permissible tools for AI development, as some coding platforms incorporate open-source components that may impose licensing obligations or encumbrances.

Data provenance is another critical area. Companies must track the origin of datasets used for training, validation, and testing to mitigate copyright infringement risks, especially amid ongoing litigation concerning unauthorized use of copyrighted materials in AI training.

The AI policy should establish mechanisms to avoid third-party IP infringement, including conducting IP landscape assessments to identify potential obstacles and whitespace opportunities for proprietary IP development. It should also define procedures for responding to infringement allegations and outline strategies for licensing third-party rights or commercializing AI assets.

While these guidelines provide a framework, companies must address industry-specific and use case-specific IP challenges unique to their AI applications. A well-crafted AI policy equips employees responsible for AI strategy implementation with clarity and direction, ultimately enhancing the company’s ability to protect and capitalize on valuable AI-related intangible assets.

Related IIPLA reading

• Latest IP news
• IP resources library
• IP guide
• IIPLA events and webinars