CHAPTER 11: AI AND INTELLECTUAL PROPERTY
Artificial Intelligence is not just reshaping industries; it’s rewriting the rules of intellectual property. Imagine a world where machines hold the pen on creativity, challenging our notions of originality and ownership. This isn't science fiction—it's our new reality.
As AI blurs the lines between human and machine creativity, legal professionals must navigate uncharted waters to protect and define the rights of all creators. This chapter discusses the dynamic intersection of AI and IP rights, uncovering the urgent questions and evolving landscapes that lawyers, firms and clients must master and consider to stay protected.
Introduction: The AI Revolution in Intellectual Property
As AI evolves, so too must our understanding of how it interacts with IP rights—copyrights, patents, trademarks, and trade secrets. This section provides an overview of the key issues and considerations that I believe legal professionals must grasp to stay ahead in this rapidly changing field.
For starters, it appears that the quick development of AI technologies has outpaced traditional IP frameworks, necessitating a reevaluation of legal doctrines to address the unique characteristics of AI-generated works. Legal professionals must consider how AI's ability to process and generate content from vast datasets impacts the traditional notions of originality and authorship. Additionally, the role of AI in augmenting human creativity raises questions about “human authorship” and joint authorship and the division of IP rights between humans and machines.
As AI continues to evolve, it is crucial for legal practitioners to stay informed about emerging technologies and their potential to disrupt established IP norms. This proactive approach will enable lawyers to provide comprehensive advice to clients navigating the complexities of IP protection in the AI era.
Copyright in the Age of AI
One of the most pressing questions surrounding AI and copyright law is whether AI-generated content can be copyrighted. Traditionally, copyright law protects original works of authorship fixed in a tangible medium of expression. However, as AI systems generate increasingly sophisticated creative outputs—from written articles to music compositions—the question of authorship becomes blurred.
The traditional concept of authorship is rooted in human creativity and intellectual labor, which poses a challenge when attributing authorship to AI-generated works. The recent decisions U.S. Copyright Office underscore the complexity of this issue.
These developments highlight the need for clear guidelines on the extent of human involvement required for AI-generated works to qualify for copyright protection. Furthermore, the distinction between AI as a tool that assists human creators and AI as an autonomous creator needs to be clarified. As AI technologies become more advanced, the legal profession must grapple with these questions to ensure that copyright law evolves in a manner that balances innovation with the protection of human creativity.
Hypothetical Scenario:
Imagine an AI-powered algorithm that autonomously generates blog posts for a legal blog. Each post is well-researched and tailored to engage readers effectively. Can the posts be copyrighted and if so, who owns the copyright to these posts—the AI developer, the user of the AI tool (lawyer or firm), or the AI itself? This scenario illustrates the complexity of determining authorship and ownership in the AI era.
Here's some additional food for thought on the AI copyright issue.
The Rothken Proposal
In his article “Copyright Liability On LLMs Should Mostly Fall On The Prompter, Not The Service,” technology lawyer attorney Ira Rothken suggests that copyright liability for outputs generated by large language models (LLMs) like ChatGPT should primarily fall on the users who provide the prompts rather than the developers of the AI systems. The proposed "Training And Output" (TAO) Doctrine suggests that, similar to the Sony Doctrine for dual-use technologies, AI developers should be protected from liability if the AI is trained on copyrighted works in a manner that does not replicate the originals but uses them to develop an understanding. The outputs, heavily influenced by user prompts, should make users responsible for any potential copyright infringement.
This approach acknowledges the dual-use nature of LLMs, emphasizing their substantial non-infringing uses and the transformative process of AI training, thus promoting innovation while respecting copyright laws.
US Copyright Office Decision
The Review Board of the United States Copyright Office continues to deny attempted registration applications for work that contains more than a de minimis amount of AI-generated content (see the link I shared above). This decision highlights the necessity for human authorship in copyrighted works and underscores the challenge of integrating AI-generated content within existing legal frameworks.
Update (July 16, 2024): USPTO issues AI subject matter eligibility guidance. This update includes three examples of how to apply USPTO’s subject matter eligibility guidance throughout a wide range of technologies.
Legislation
It's interesting to note that as of the writing of this book, a bipartisan group of senators has just rolled out a groundbreaking bill that's set to be a game-changer for artists, songwriters, and journalists. The Content Origin Protection and Integrity from Edited and Deepfaked Media Act (COPIED Act) is designed to safeguard creators from having their work exploited by AI models without permission. This bill doesn't just stop at protecting original content creators—it also introduces clear measures to help us all spot AI-generated material. By tackling the rise of harmful deepfakes, the COPIED Act aims to bring more transparency and trust back to the digital media landscape. Here's a good article if you're interested in reading more.
Patents: AI as an Inventor?
Patent law traditionally requires human inventors to be named. However, advancements in AI raise intriguing questions about inventorship.
Can an AI system be credited as an inventor? What are the implications for patentability and the rights of inventors? This section explores recent legal and policy developments shaping the landscape of AI and patent law.
The debate over AI inventorship is far from theoretical; it has real-world implications for patent applications and the innovation ecosystem. Recent cases, such as Thaler's DABUS (Device for the Autonomous Bootstrapping of Unified Sentience) applications, have brought this issue to the forefront, challenging patent offices worldwide to reconsider the definition of "inventor."
Thaler's DABUS applications refer to patent filings that list an AI system, DABUS, as the inventor. These applications, spearheaded by Dr. Stephen Thaler, have sparked significant legal and philosophical debates about the nature of inventorship and the role of AI in the creative process. While some jurisdictions, like the UK and the EU, have upheld the requirement for human inventors, others are still grappling with the implications of AI's role in the inventive process.
Additionally, AI's capability to rapidly analyze data and generate innovative solutions poses a challenge to the traditional patent examination process, which must adapt to evaluate the novelty and non-obviousness of AI-generated inventions. Legal professionals must stay attuned to these developments and consider the potential for AI to not only assist in but also independently drive innovation, influencing both patent strategy and policy.
If you'd like to learn more information about AI and patents, I invite you to read "Can AI Hold Patents According to the US Patent Office?" shared by my friend, patent lawyer, and aerospace engineer (pretty darn impressive if you ask me), Karima Fathi Gulick.
Trademarks and Brand Protection: Leveraging AI for Enhanced IP Strategy
AI offers powerful tools for brand protection and enforcement. From automated trademark monitoring to predictive analytics for brand strategy, AI empowers legal professionals to safeguard their clients' brands more efficiently than ever before.
The integration of AI into trademark law practice goes beyond mere automation; it enables more strategic and proactive brand management. AI systems can continuously scan for potential trademark infringements across digital platforms, providing real-time alerts and allowing for swift enforcement actions. Furthermore, AI-driven analytics can offer deeper insights into market trends and consumer behavior, aiding in the development of more effective brand protection strategies.
By predicting potential challenges and identifying opportunities for brand expansion, AI helps legal professionals advise clients on optimizing their trademark portfolios. Embracing these technologies not only enhances the efficiency of trademark monitoring and enforcement but also provides a competitive edge in maintaining and growing brand value in an increasingly digital marketplace.
Trade Secrets and AI: Protecting Confidential Information in a Data-Driven World
AI-driven data analytics can enhance the identification and protection of trade secrets, yet they also raise concerns about data security and confidentiality. This section discusses strategies for leveraging AI to safeguard trade secrets while navigating legal and ethical considerations.
AI can significantly bolster trade secret protection by identifying patterns and anomalies that may indicate unauthorized access or leaks, thus enabling quicker responses to potential threats. Additionally, AI can assist in classifying and managing sensitive information, ensuring that trade secrets are adequately protected within an organization's data infrastructure.
However, the use of AI also introduces new risks, such as the potential for AI systems to inadvertently expose sensitive information or the ethical dilemmas surrounding the extent of surveillance necessary to protect trade secrets. Legal professionals must balance these considerations, implementing robust AI-driven security measures while maintaining compliance with data privacy regulations and ethical standards.
Regulatory and Ethical Considerations: Navigating the Complexities of AI and IP Law
As AI technologies proliferate, so too do regulatory and ethical challenges. From data privacy concerns to algorithmic bias, legal professionals must navigate a complex landscape of laws and guidelines governing AI applications. This section explores the regulatory framework and ethical principles that shape AI and IP law, providing guidance on compliance and best practices.
The global nature of AI development adds another layer of complexity, as legal professionals must consider varying international regulations and standards. For instance, the European Union's General Data Protection Regulation (GDPR) imposes stringent requirements on data handling and privacy, influencing how AI systems are designed and deployed.
Similarly, the AI Act proposed by the European Commission aims to establish comprehensive rules for AI development and usage, emphasizing transparency, accountability, and human oversight. In the United States, regulatory bodies like the Federal Trade Commission (FTC) are increasingly scrutinizing AI practices to prevent unfair or deceptive practices.
Ethical considerations also play a critical role in shaping AI's integration into IP law. Issues such as ensuring equitable access to AI technologies, preventing discriminatory outcomes, and maintaining the integrity of creative works demand careful attention. I discussed the ethical implications in detail in an earlier chapter but once again, it’s important to remember that legal professionals must advocate for ethical AI practices within their organizations and for their clients, promoting the development of AI systems that respect human rights and societal values.
Furthermore, the establishment of AI ethics committees or advisory boards can help organizations navigate these challenges, providing oversight and guidance on ethical dilemmas. By fostering a culture of ethical AI development and use, legal professionals can help build public trust in AI technologies, ensuring that their benefits are realized while mitigating potential harms.
The Future? Blockchain and Web3: Revolutionizing IP Applications and Enforcement
Web3 technologies, particularly decentralized blockchain and smart contracts, offer transformative potential for the application and oversight of IP rights. A blockchain is a decentralized digital ledger that records transactions across many computers in a way that ensures the data cannot be altered retroactively. This technology's qualities—such as immutability, transparency, and decentralization—make it particularly suited for managing IP rights.
Ownership of IP assets can be recorded on a blockchain, providing a transparent and unchangeable record of provenance that ensures the originality and authenticity of creative works. This transparency can significantly reduce disputes over IP ownership and provenance, as all transactions and transfers are publicly verifiable.
Blockchain’s transparency ensures that every transaction related to an IP asset—whether it is a sale, transfer, or usage agreement—is recorded in a way that is publicly accessible and immutable. This creates a clear chain of custody and proof of ownership that cannot be tampered with, thus significantly enhancing the reliability of IP rights management.
For instance, when an artist creates a digital artwork, they can mint it as a non-fungible token on most blockchains or inscribe it as an Ordinal on a bitcoin blockchain. Although different in technology, a NFT is hashed to a third-party servier and an Ordinal is inscribed and becomes part of the bitcoin blockchain, for purposes of this chapter I’m going to refer to both as a “digital artifact.”).
A digital artifact is a unique digital asset that represents ownership of a specific item, such as a piece of art, a music track, or even a video clip. Each digital artifact has a unique identifier that distinguishes it from other tokens, ensuring that the item it represents is one-of-a-kind. This digital artifact acts as a digital certificate of authenticity and ownership. Each time the artwork is sold or transferred, the transaction is recorded on the blockchain, providing a permanent and transparent history of ownership.
Smart contracts, which are self-executing contracts with the terms of the agreement directly written into code, can automate many aspects of IP enforcement. For example, a musician could encode a smart contract to automatically distribute royalties every time their song is streamed or purchased. This eliminates the need for intermediaries, reduces administrative costs, and ensures that creators are paid promptly and fairly.
The smart contract can specify terms such as the percentage of royalties each stakeholder receives and the conditions under which payments are made. Once the conditions are met—such as the streaming of the song—the smart contract automatically executes the payment.
Furthermore, AI integration with blockchain can significantly enhance IP monitoring and enforcement. AI algorithms can be designed to continuously scan the blockchain for transactions and activities related to specific IP assets. For instance, if an unauthorized transfer of a copyrighted digital artwork is detected, AI can flag the activity and initiate enforcement actions. These actions might include alerting the rightful owner, generating a cease-and-desist notice, or automatically revoking access to the infringing party. This proactive approach ensures that IP violations are addressed promptly, minimizing the potential damage to the rights holder.
Example:
Imagine a fashion designer who creates a unique digital design and uploads it to a blockchain as an digital artifact. This digital artifact includes a smart contract that specifies the terms of use, sale, and royalty payments. If another designer wants to use this digital design for a limited edition clothing line, they can purchase a license through the blockchain. The smart contract automatically enforces the terms of the license, such as the number of units that can be produced and the duration of the license. Each time a clothing item featuring the design is sold, the smart contract automatically distributes royalties to the original designer.
Alternatively, the designer could use Bitcoin Ordinals, a similar technology that inscribes data onto individual satoshis, the smallest unit of Bitcoin. Like NFTs, Bitcoin Ordinals provide a way to create unique digital assets on the Bitcoin blockchain. The designer inscribes their digital design onto a satoshi, creating a unique and verifiable record of ownership. Transactions involving this design are recorded on the Bitcoin blockchain, providing the same benefits of transparency and immutability.
In both cases, if someone attempts to use the design without authorization, AI integrated with the blockchain continuously monitors for such infringements. When unauthorized use is detected, the AI flags the transaction and triggers the smart contract to take action. This might involve sending a notification to the infringer, documenting the infringement on the blockchain, and restricting access to the design. The designer is immediately alerted and can take further legal action if necessary.
This example illustrates the seamless integration of blockchain, smart contracts, and AI in protecting and enforcing IP rights. The blockchain provides a transparent and immutable record of ownership and transactions, the smart contract automates enforcement of IP terms, and AI ensures proactive monitoring and prompt response to potential infringements.
This combination offers a powerful toolset for IP law, enhancing both the protection of creators' rights and the efficiency of IP management. By leveraging these technologies, legal professionals can be faster, better and provide more robust and responsive services to their clients, ensuring that IP rights are safeguarded in an increasingly digital and decentralized world.
Real-World Examples
My dear friend, IP lawyer Pete Salsich, III, covered several of these issues in recent episodes on The Screen Lawyer Podcast. In this episode titled, “AI, Copyright, and the Future of Music,” he and Gary Pierson unpacked the groundbreaking lawsuit brought by major music labels against AI music innovators Suno and Udio over copyright infringement.
I also found Pete’s episode, about OpenAI allegedly using Scarlett Johansson’s voice without her permission entertaining and educational. In my opinion, Pete's analysis applies to most “right to publicity” related AI content claims.
The horizon of AI and intellectual property is teeming with potential and challenge. We've explored how AI is reshaping copyright, patents, trademarks, and trade secrets, pushing the boundaries of traditional IP law. The question of authorship, the role of AI as an inventor, and the innovative use of blockchain and smart contracts are just a few of the issues that will define the future of legal practice.
But as we stand on this exciting frontier, there's another crucial aspect we must address: protecting the privacy and security of the data that fuels these AI systems. How do we safeguard sensitive information in this rapidly evolving digital landscape? The answers await in the next chapter, where we'll dive into the critical measures and ethical considerations essential for privacy and security in the age of AI. Stay tuned—this is where the journey truly begins.