Author : Prachi from Galgotias university Co-Author : Naina from Galgotias university
Introduction: The Dematerialization of Value
A structural reordering of how human societies create, capture, and distribute value, from industrial to digital economy, represents an irreversible shift in the economic foundations of societies. For a major corporation, power during the 20th century was almost entirely in its physical assets: real estate, factories and machinery, natural resources, etc. Now, corporate value has been radically ‘dematerialized’ by a “focus on intangible assets” in the digital economy such as proprietary software, algorithms, databases, brand-value and networks. As the World Intellectual Property Organization notes, investments in intangible assets have tripled investments in physical assets over the last two decades, creating an economic environment in which digital data and code constitute the basic architecture of global trade. This restructuring of society is an existential crisis for legal regulation. Traditionally the law regarding intellectual property, which sought to regulate physical objects like books, mechanical devices, and goods, now needs to systematically reinterpret itself: traditional demarcations between patents, copyrights, trade secrets and trade marks are blurring. Digital objects occupy a paradoxical middle space; a software platform requires intellectual property protection both for its literally presented source code (under copyright) and for the underlying procedures in its code (under patent) as well as for the proprietary algorithms, and datasets developed for machine learning. Students reading law need to comprehend how these competing values are organized, so beyond mere black-letter law, students should appreciate structural tensions such as those between statutory territoriality versus actual free flow of digital data, and those between encouraging private creativity versus maintaining the public digital commons, not to mention the dilemma of applying age-old concepts to such entities as AI. I here give a comprehensive legal analysis of these changes to IP law.
Copyright Law in the Digital Era
The copyright system has been the most volatile battleground in the digital transformation, as fundamental doctrines established in an age of physical media- fixity, authorship, and infringement- are challenged by the Internet and by generative AI.
- The Demise of the First Sale Doctrine and the Birth of Licensing
In the physical economy, copyright law maintained an appropriate balance between the copyright holder’s rights and the public’s access to material goods through the First Sale Doctrine, codified in 17 U.S.C. § 109 in the United States and internationally through the concept of the Exhaustion of Rights. It stated that once a copyright holder sells a lawful physical copy of a work (for instance, a book or a vinyl record), their right to control the further distribution of that specific copy ends. A subsequent purchaser can freely resell, lend, or donate it.
In the digital economy, the doctrine is essentially nonexistent. When a customer “purchases” an ebook, software or game on a digital platform, what they are truly buying is not a product, but rather an End-User License Agreement (EULA). These end-user license agreements are, fundamentally, non-transferable and revokable licenses to access content stored on remote servers.
The Supreme Court’s decision in Vernor v. Autodesk, Inc. Established that software purchasers are licensees rather than owners if the copyright holder restricts the ability to transfer and imposes limitations on use. This made the First Sale Doctrine irrelevant in the context of digital access. It allowed copyright holders to kill the used market for digital items, while giving them the ability to exert perpetual upstream control over distribution.
- The Fair Use Doctrine and Training AI
The widespread use of generative artificial intelligence have placed the doctrine of Fair Use (or fair dealing in civil law regimes) at the forefront of copyright debates. This type of AI requires billions of images, texts and codes scanned from the internet as training sets and has spurred major lawsuits by artists, authors, and news publishers arguing that copying and using their copyrighted work in training sets is massive copyright infringement.
Under US copyright law, whether or not this constitutes Fair Use is governed by a four-factor test
1. The purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
2. The nature of the copyrighted work;
3. The amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
4. The effect of the use upon the potential market for or value of the copyrighted work.
AI developers have relied heavily on the ruling in Authors Guild v. Google, Inc., in which the Second Circuit found that Google’s mass scanning of millions of books was a transformative fair use that created an entirely new utility and market not competing with the books themselves.
However, creators have countered that AIs are not creating indices; they are producing competing content. Courts will also likely look at the Supreme Court’s decision in Andy Warhol Foundation for the Visual Arts, Inc. V. Goldsmith, in which they narrowed the definition of “transformative use” by stating that a work is not transformative if its purpose is similar to the purpose of the original work, especially if there’s a commercial component. Because AIs can produce text, art, and code that compete directly with human-created content using the data they are trained on, courts are investigating whether the ingestion stage can be properly divorced from the commercial output phase.
- The AI Authorship Problem
The flip side of training is output: is a work created by an AI copyrightable?
Historically, copyright systems globally have adhered to an unwritten but absolute rule that copyright must be in a human creation. The Supreme Court case Burrow-Giles Lithographic Co. V. Sarony identified copyright as the exclusive right granted to an “author” over their “writings” and “compositions” which implicitly suggested “an original intellectual production of an author” and also that authors must be human. This precedent was reinforced in Feist Publications, Inc. V. Rural Telephone Service Co., which affirmed that for a work to be copyrightable, “creativity is the touchstone of copyright protection, not merely the sweat of the brow.”
Autonomous systems are now directly challenging these concepts. In Thaler v. Perlmutter, the U.S. District Court for the District of Columbia upheld the refusal by the U.S. Copyright Office to register a piece of art created autonomously by an AI system, solidifying the notion that authorship must originate from human. The U.S. Copyright Office has even gone on to issue guidelines requiring that applicants identify AI-generated material, while also stating that “because the AI ‘artist’ did not create the work autonomously but responded to and interacted with the prompt…the human’s creative work was part of the overall system, enabling it to function.” Thus, while the specific arrangement or prompting that leads to a piece generated by an AI is protectable (by the human author), the generative content itself seems to exist in the public domain.
3. Patent Law and Software-Driven Systems
Patent law was designed to protect tangible inventions with a functional or utilitarian purpose. As technology shifts the economy toward digital concepts, this framework will be frequently updated in order to deal with the abstract logic, architecture, and machine learning systems used.
Section 101 Jurisprudence and Software Patent Eligibility
Digital innovations’ primary hurdle is the law of patent eligibility. In the US, 35 U.S.C. §101 permits the granting of patents for any “new and useful process, machine, manufacture, or composition of matter”.Courts long recognized 3 judicial exceptions to this grant of patents for laws of nature, natural phenomena, and abstract ideas.
Per step 1, courts examine whether the claim is directed toward an abstract idea. If so, they enter step 2 where the courts consider whether individual claim elements, or the claim taken as a whole, adds to the claim an “inventive concept” transforming it into a patent-eligible application; simply computer elements operating in a conventional manner add nothing.
Post-Alice jurisprudence resulted in a severe cutback in pure software and fintech patenting; now, only technological improvements on the computer itself (e.g., an advanced algorithm or an optimized memory system) and not merely business methods executed by a computer will pass a Section 101 test.
AI as an Inventor
As copyright law grappled with questions about an author’s authorship by an autonomous system, so has patent law faced the question of whether AI systems can be named inventors. The debate played out in courts worldwide through Dr. Stephen Thaler’s attempts to list an AI system, DABUS, as the sole inventor for a food container and emergency beacon.
Legal consensus has uniformly rejected the notion that AI systems can be inventors:
US: In Thaler v. Vidal, the Federal Circuit decided based on the text of the Patent Act; the law defines an inventor as an “individual” in 35 U.S.C. §100(f) and the Supreme Court had already held that an “individual” has to be a natural person to satisfy this definition.
EPO & UK: similar conclusion was reached in the UK Supreme Court;[^15] a court ruled that under the UK’s Patent Act of 1977, an inventor is necessarily a natural person. Additionally, the EPO Guidelines for Examination state that the applicant must list a human being as the inventor as well, and this reflects an international view of patents as an incentive for human thought and investment rather than the calculations of an automaton as a distinct legal entity.
4. Trademarks, E-Commerce, and Brand Identity
Trademarks safeguard consumers’ brand identity and prevent confusion in the market. In the digital era, the concept of a “marketplace” has moved from a physical brick-and-mortar storefront to algorithm-driven e-commerce and virtual worlds.
Intermediary Liability for Online Marketplaces
Arguably, the most important digital trademark issue has been establishing when an e-commerce platform can be held liable for third-party sellers’ sale of counterfeits.
The seminal case setting up the foundation of secondary trademark liability is Tiffany (NJ) Inc. V. EBay Inc.. In that case, Tiffany sued eBay for contributory trademark infringement, asserting that the site had actual knowledge of counterfeit Tiffany goods on its site and profited from them.
The Second Circuit ruled for eBay on the grounds that the owner of an on-line platform cannot be contributorily liable for trademark infringement with only the knowledge that infringing materials exist on the site; rather, actual liability arises only when the owner has specific and contemporaneous knowledge of particular listings which infringe, yet takes no action to remedy the same.
The situation has been complicated by the integration of logistics services, such as the “Fulfillment by Amazon” program. When a third party, on-line intermediary not only hosts thelistingbut also stores, picks, packs, and ships an infringing item-while simultaneously processing the transaction-it increases the risk of liability.
In Amazon.com Inc. V. Louboutin, the Court of Justice of the EU ruled that even an online intermediary may be liable for trademark infringement where an average reasonably informed internet user may perceive that the operator of the site itself is offering and selling the counterfeit products under its brand image.
Legal Framework / Case Jurisdictional Scope Primary Legal Standard for Intermediary Liability
Tiffany v. EBay(Second Circuit) United States Requires contemporaneous, specific knowledge of actual infringing activity; generalized awareness of potential counterfeits does not create liability.
DMCA(Digital Millennium Copyright Act) United States Copyright law primarily, but requires active, formal notice-and-takedown architecture.
Amazon v. Louboutin(CJEU) European Union An online intermediary can be held directly liable if an average, reasonably well-informed Internet user believes that the platform is the active seller of the third party counterfeit product.
Keyword Advertising and SEO
The blurring of brand identities and search algorithms have resulted in other trademark disputes, especially involving keyword advertising such as Google Ads. Companies routinely bid on their competitor’s trademarked names as search terms to ensure that searches on a certain brand name pull a link to the competitor’s website at the top of search results.
Trademark owners first alleged unauthorized commercial use of their marks in such advertising. The courts have predominantly concluded that such usage is lawful if the consumer is not likely to be confused about the origin of the products or services offered. Using a multi-factor test to analyze the likelihood of confusion, courts have held in Google LLC v. Interflora Inc. that trademark usage in keyword bidding itself is not per se infringing. Infringement exists only when the advertisement itself, and not merely the fact that it is displayed as a result of trademarked keyword usage, is likely to confuse the consumer and not make it reasonably apparent that there is some nexus to the trademark owner.
5. Trade Secrets and Data Governance
With the narrower patentability of software under Section 101, organizations are increasingly seeking trade secret law protection for their more sensitive digital assets. Algorithms, database schemas, and raw data fit neatly under the definition of a trade secret and can be kept hidden indefinitely.
The Framework for Trade Secret Protection
Unlike patents and copyright, trade secret rights do not require any government filing. Trade secrets are uniformly covered by the Defend Trade Secrets Act in the United States , and by state-level adoption of the Uniform Trade Secrets Act. The EU Trade Secrets Directive provides a harmonized framework for the EU.
In order for an asset to fall under trade secret law, it must satisfy three statutory tests:
1. The asset must derive economic value from not being generally known;
2. The asset must not be readily ascertainable through proper means by others; and
3. The trade secret must be subject to reasonable measures under the circumstances to maintain secrecy.
In the digital world, reasonable measures means much more than simply locking away a physical document. An active digital architecture must exist in the form of strong end-to-end encryption protocol for data, restrictive and granular role-based database permissions, enforced employment non-disclosure agreements, ongoing monitoring systems to detect data scrapers/exfiltrators, and so forth. In the absence of standard cybersecurity safeguards, courts will often deem the data to no longer constitute a protectable trade secret .
Data Scraps, Web Scraping and Misappropriation
Because data has real value, web scraping has become a common cause of action. In practice, it refers to automated programs that scan public websites and extract vast amounts of data and send it back for use in a private analytical program or for training an AI model. Legality of this behavior is highly contested as it sits at the intersection of trade secrets, contract law, and computer crime laws.
In hiQ Labs, Inc. V. LinkedIn Corp. , the Ninth Circuit ruled that scraping public web data did not violate the Computer Fraud and Abuse Act since it was publicly accessible to anyone with an internet connection, therefore it was not a “unauthorized access” of a protected computer.
Consequently, data owners have shifted their defensive strategy away from criminal statutes towards breach of Contract claims, through embedding explicit anti-scraping language in their ToS, so that they can sue the scraper for breach of contract, regardless of whether the data is public. However, if the extracted data can be structured and packaged into a non-public format, scrapers may also be found to be guilty of trade secret misappropriation if they used inappropriate or deceptive measures to bypass the electronic security
6. The Intersection of Anti-Trust, Privacy, and Intellectual Property
In the digital economy, a phenomenon called “network effects” determines success: platforms are more useful and powerful with more users. This creates a self-defeating scale and ultimately dominant, or “gatekeeper,” ecosystems where intellectual property rights, antitrust law and data privacy mandates frequently conflict.
IP Monopolies vs. Antitrust Interoperability
By definition, an intellectual property right provides its owner with an exclusive legal monopoly over a particular asset. In normal markets this monopoly is limited by the availability of competing or substitute goods. However, in digital platform markets, network effects ensure a single IP asset becomes a critical economic bottleneck.
This is most notably illustrated in the debate over Application Programming Interfaces (APIs) and data interoperability. APIs are essentially software interfaces which are used to allow separate programs to communicate with and share information with each other. If a dominant platform can use either copyright or patent law to entirely prevent a competitor from integrating with the platform’s API, it can block secondary competition across entire digital markets.
Google LLC v. Oracle America, Inc. raised the issue of copyright infringement when the Supreme Court examined whether Google’s creation of the Android operating system by replicating approximately 11,500 lines of the declaring code of Oracle’s Java API amounted to copyright infringement. The Court concluded that Google’s use was fair use as a matter of law.
In its ruling the majority clarified that declaring code is dramatically different to the core functional code as the former essentially acts as an organisational framework providing the mechanism for programmers to invoke pre-existing functional code. Therefore the inability to use these interface lines would act as a legal lock-in hindering the development of subsequent software and undermining the constitutional purpose of IP rights.
Data privacy laws vs IP ownership
The rise of Comprehensive data privacy legislation such as the EU’s General Data Protection Regulation (GDPR) and the Californian Consumer Privacy Act (CCPA) directly conflicts with the current concept of IP ownership. While intellectual property law asserts that a business designing an analytical database holds the copyright of this compilation, and the proprietary insights gained through that analytical database as trade secrets, data privacy laws provide the individual consumer with comprehensive rights over their personal data. These rights include:
The Right to Deletion (the “Right to be Forgotten”): Consumers can require businesses to remove personal data points about themselves from the database.
The Right to Data Portability: Consumers can require that a business transfer any structured personal data pertaining to them to a commercial competitor.
This infringes on an enterprise’s IP assets, such as trade secrets and algorithm investments. For instance, an AI model trained using personal data would have to be retrained if the consumer demands that their data be deleted, requiring a significant investment in capital to compensate for the lost training data.
7. Comparative Analysis: Global Legislative Responses
As the economy crosses national borders, national legal regimes need to adjust their statutory frameworks accordingly. This section considers major legislative responses enacted by the United States, European Union and India.
The United States: Judicial Adaptation and the DMCA
In the United States, judicial interpretation of existing statutes prevails, along with specific foundational acts like the DMCA.512U.S.C. § 512 grants a copyright-infringement “safe harbor” to online intermediaries if they have a working notice and takedown system in place.
However, this architecture, designed for static web-hosts, is tested by the automated distribution of content through algorithms. To supplement this approach, federal agencies have stepped in with direct administrative guidance :USPTO Inventorship Guidance, clarifying that AI cannot be named as inventor but that human inventors can obtain patents for works produced with assistance of AI, if the human contribution was pivotal to the invention’s conception.Copyright Office Frameworks, auditing registration requests for the disclosure of AI components, ensuring a baseline for human authorship registration.
The European Union: Direct Statutory Intervention
In direct contrast to the US, the EU has adopted a proactive, codifying legislative approach, enacting overarching digital legislation designed to realign power between rights-holders and technology companies.
The EU AI Act: The first comprehensive, horizontal law specifically aimed at AI.514 The Act establishes stringent transparency requirements for general-purpose AI models, specifically demanding they document and publicize a clear summary of the copyright-protected materials used for training their AI systems.
The DSA & DMA: These acts fundamentally overhaul intermediary liability rules, moving away from safe harbor provisions and toward mandated due-diligence requirements for online platforms.515 For “gatekeeper” dominant platforms, the EU’s anti-trust provisions ban practices such as self-preferencing their own services or mandating closed API architecture for proprietary interoperability.
The EU Copyright Directive (Art. 17): Pushes liability for user-uploaded copyright infringements squarely onto large platforms rather than creators. Platforms are liable for copyright violations unless they take proactive steps by using advanced filtering technologies to block infringing content before it becomes publicly available.516
India: Embracing Digital Growth through Pragmatic Frameworks
India has structured its IP laws to nurture its rapidly growing digital ecosystem, enabling innovation while promoting public access.
The Digital Personal Data Protection Act: This legislation establishes a statutory regime for personal data, outlining the individual’s rights over her data and defining the way corporate entities compile this data or protect their proprietary trade secret algorithms.517
Sec. 3(k) of the Patents Act: This provision ensures that software does not become patentable. Specifically excluding computer programs per se and mathematical methods from patent protection, the Act and the Indian Patent Office require software inventions to have a clear, tangible technical effect on, or interface with, the physical structure of a machine for it to be patentable.518
Copyright Act and Algorithmic Enforcement: Indian courts are a leading forum for the dynamic injunction process. In cases of widespread digital piracy, Indian courts have issued ‘John Doe’ (Ashok Kumar) orders, compelling internet service providers to instantly block mirror and proxy websites in order to protect digital media assets.
8. Conclusion: The Future of Digital Intellectual Property
Digital economy issues have irrevocably disrupted established ip law principles and rules. Tangible assets were once the primary economic value producers but intangible ones like software, data and algorithms now rule the world. The framework of the law of products needs to adjust to one of licensing, algorithms and automatic data analysis. The next generation of lawyers needs to be well-versed in this multidisciplinary field, which now sits hand-in-hand with contract law, as represented by sticky EULAs, cyber-security (trade secret compliance) and public policy (AI governance, privacy), among others. As these technologies advance, the key challenge for legislatures, judges and attorneys, respectively, will continue to be to engineer an impartial system for the reward of human innovation and the protection of business investments while maintaining the public digital space as an open, competitive commons.