A technology company based in Buenos Aires deploys a generative AI tool to produce marketing copy, product imagery, and software components at scale. The output is commercially valuable. The company registers some of it, licenses some, and treats the rest as proprietary. Then a competitor begins using near-identical material. When the company's legal team reaches for Argentina's intellectual property legislation, they find that the statutes were written with a human author firmly in mind. The question of who – or what – created the work becomes the central and unresolved issue.
AI-generated works in Argentina occupy a legal grey zone under current intellectual property legislation, which conditions copyright protection on human authorship and original creative expression. No dedicated statute governs AI-generated output, and Argentine courts have not yet produced a settled body of decisions on the matter. Businesses operating in this space must structure their IP position proactively, because the default rules were not designed with autonomous AI creation in mind.
This analysis examines the doctrinal foundations, competing interpretations, the gap between statute and practice, cross-border dimensions for Americas clients, and the strategic steps available today. It also considers how global developments – including AI Act compliance discourse in Europe – are beginning to influence the debate in Argentina.
Doctrinal foundations: what Argentine intellectual property law says about authorship
Argentina's intellectual property regime rests on legislation that dates, in its core structure, to the early twentieth century. The central statute governing copyright – Argentina's intellectual property law – defines the author as the natural person who creates a work. This definition was drafted without any contemplation of software liability issues or autonomous machine creativity. It reflects a civil law tradition in which the author's personality is inseparable from the protected work.
Under this doctrinal model, copyright protection arises automatically upon the creation of an original work by a human author. Registration with the Dirección Nacional del Derecho de Autor (National Copyright Directorate) is declaratory, not constitutive. It creates a presumption of authorship and a fixed date, but it does not bring the work into existence as a protected object. The work's protectability depends entirely on the presence of human creative choices.
The originality threshold in Argentine copyright doctrine is modest compared to some other civil law systems. Courts have generally held that a work is original if it reflects the author's personal intellectual effort. a standard sometimes described as reflecting the author's "stamp." This low threshold has historically allowed a wide range of works to receive protection. The difficulty with AI-generated output is not the height of the originality bar. It is the prior question: whether any human personality is present in the creation at all.
Software receives protection under intellectual property legislation as a literary work. This classification has been significant for technology licensing arrangements and for the software liability questions that arise when programs malfunction. However, the classification of software as a literary work also imports all the authorship conditions that apply to literature. An AI model that writes code does not automatically satisfy those conditions merely because code is categorised as a literary work.
Patent law in Argentina imposes separate conditions. Patentable inventions must be novel, involve an inventive step, and be capable of industrial application. Argentine patent legislation excludes from patentability certain categories of abstract subject matter, including mathematical methods and mental processes. Whether an AI-generated technical solution constitutes a patentable invention – and who the inventor would be – raises questions that Argentine patent doctrine has not yet definitively answered. The inventor must, by current doctrine, be a natural person.
The authorship gap: competing interpretations and where courts are heading
The central doctrinal problem is what practitioners call the authorship gap. Existing legislation requires a human author. Sophisticated AI tools produce output with minimal human creative input at the generative stage. Three competing interpretations have emerged in Argentine legal scholarship and early judicial reasoning.
The first interpretation applies the legislation strictly. Under this view, output produced autonomously by an AI system without human creative choices at the point of generation is not protected by copyright. The human who operates the tool makes technical choices. selecting the model, entering prompts, choosing output parameters. but these choices are not, on this analysis, the kind of creative expression that intellectual property legislation protects. The work enters the public domain immediately upon creation.
The second interpretation extends existing doctrine by analogy. Argentine copyright law has long recognised works of collective authorship, works produced by employees within the scope of employment, and works commissioned under contract. Proponents of the analogical approach argue that the human or company who directs, trains, and deploys the AI system exercises sufficient creative control to qualify as author or right-holder. The AI is treated as a sophisticated tool – analogous to a camera or a typesetting machine – and the human behind it is the author.
The third interpretation holds that Argentine legislation contains a genuine gap that only the legislature can fill. On this view, neither strict application nor analogy produces satisfactory outcomes across all fact patterns. The legislator must create a new category. perhaps a sui generis (unique. Purpose-built) right akin to the database right in European intellectual property legislation. that protects investment in AI-generated output without making claims about authorship in the traditional sense.
Argentine courts have not yet produced a consolidated line of decisions on this precise question. The few cases that have touched on AI-related IP issues have generally involved human-assisted AI output rather than fully autonomous generation. In those cases, courts have tended to apply the tool-analogy approach, treating the human operator as the author. Whether that reasoning extends to outputs where the AI's generative contribution overwhelmingly dominates is an open question.
Practitioners in Argentina note a significant risk of inconsistency. Different chambers of the Cámara Nacional de Apelaciones en lo Civil (National Civil Court of Appeals) may reach different conclusions on analogous facts before the Corte Suprema de Justicia de la Nación (Supreme Court of Argentina) has the opportunity to settle the matter. In the interim, businesses face genuine uncertainty about whether their AI-generated assets are protected.
The algorithmic accountability dimension adds complexity. When an AI system produces output that infringes third-party rights – reproducing copyrighted training data, for example – the question of who bears software liability is equally unsettled. The developer of the model, the entity that deployed it, or the end user may each face claims depending on how Argentine civil liability doctrine is applied. Courts have traditionally required fault or causation attributable to a legal person. Mapping that requirement onto an AI pipeline involves interpretive choices that courts are only beginning to confront.
Statute and practice: where the gap is widest and what businesses are doing
The gap between Argentina's intellectual property statutes and actual commercial practice is most visible in three areas: content production, software development, and technology licensing.
In content production, businesses routinely use AI tools to generate text, images, audio, and video at volumes that no human workforce could match. Many of these businesses apply for copyright registration with the National Copyright Directorate, listing a human author – typically the prompt engineer, the creative director, or the commissioning company. The Directorate has not publicly established a clear policy on AI-generated works. Registration applications that disclose AI involvement have sometimes been accepted and sometimes returned for clarification. The absence of a published policy creates practical unpredictability.
In software development, AI-assisted coding tools now generate substantial portions of production codebases. The employment contracts and contractor agreements under which human developers work typically assign IP to the employer or commissioning party. Where the AI tool's contribution to a codebase is significant, the question arises whether that assignment clause captures something the human developer never created. Argentine employment legislation governs the assignment of works created by employees, but the legislation assumes a human creator. A careful reading of those provisions does not straightforwardly resolve the AI-generated component question.
Technology licensing presents a third pressure point. A licensor granting rights over a software product that contains AI-generated components must represent that it holds the necessary rights. If those rights are legally uncertain – because the AI-generated components may be unprotected or may infringe training-data rights – the representation is potentially false. Warranty and indemnity provisions in technology licensing agreements across Argentine and cross-border transactions are increasingly being negotiated around this risk. Counterparties are asking for specific representations about AI involvement in the licensed product.
Digital services businesses face additional exposure under Argentina's data protection and digital services rules. When AI systems process personal data to generate personalised output, the intersection of intellectual property questions with data protection obligations creates layered compliance requirements. Businesses operating AI-driven digital services in Argentina should map their workflows against both bodies of law – not just IP legislation.
To explore how AI and technology law in Argentina applies to your specific deployment model, a tailored legal assessment is the most effective starting point.
Cross-border implications for Americas clients
Argentina sits within a regional and global IP system that adds layers of complexity to the domestic uncertainty. The country is a member of the Berne Convention, the TRIPS Agreement under the World Trade Organization framework, and other multilateral instruments that shape minimum standards for IP protection. These instruments were also drafted with human authorship in mind. They do not resolve the AI question, but they do constrain the options available to the Argentine legislature when it eventually acts.
For businesses operating across multiple Americas jurisdictions – Brazil, Mexico, Colombia, Chile, and Argentina being the most commercially active markets – the variation in approach to AI-generated IP is a material risk. Brazil has begun a legislative process to address AI regulation. Mexico's IP regime raises similar authorship questions under its civil law tradition. An AI-generated work that may be protectable in one jurisdiction under a tool-analogy doctrine may be in the public domain in another jurisdiction that applies strict human-authorship requirements.
This divergence has direct consequences for technology licensing strategy. A company licensing AI-generated content across the Americas cannot assume that a single IP ownership structure will be valid in each market. The licensor needs jurisdiction-by-jurisdiction analysis of whether the underlying rights exist and whether they are enforceable. This is not merely a compliance question. It is a commercial question: the value of the licensed asset depends on whether the licensee can actually use it without third-party interference.
The European dimension is also relevant for companies with transatlantic operations. The EU's AI Act introduces obligations around high-risk AI systems, transparency, and algorithmic accountability that go beyond intellectual property and touch on how AI outputs are disclosed and documented. Argentine companies exporting AI-generated products or services into the EU market must consider AI Act compliance requirements alongside their domestic IP position. The practical effect is that documentation and disclosure practices that satisfy EU regulatory requirements may also strengthen the evidentiary record that an Argentine court would need to apply the tool-analogy doctrine to a particular AI-generated work.
For clients managing cross-border IP portfolios that include AI-generated assets, the analysis of AI-generated works and intellectual property in the United States offers a useful comparative reference point. Particularly on the human-authorship requirement and its practical consequences for registration strategy.
Enforcement of IP rights in AI-generated works across borders raises a further set of questions. If an Argentine company seeks to enforce its claimed rights in AI-generated content against a defendant in another jurisdiction. The enforcing court will apply its own conflict-of-laws rules to determine which country's law governs the existence and scope of the right. A right that is uncertain under Argentine law may be treated differently – better or worse – by a foreign enforcing court. Structuring the creation and assignment of AI-generated works with cross-border enforceability in mind requires planning that goes beyond simply registering in each country.
For a comprehensive assessment of intellectual property protection in Argentina across both AI-generated and traditional works, contact us to discuss the specific portfolio structure that fits your business model.
Strategic recommendations and what to do now
Given the doctrinal uncertainty and the pace of commercial adoption, businesses cannot wait for legislative reform before building an IP protection strategy for AI-generated assets. Several practical approaches can materially reduce risk under the current conditions.
The first priority is documentation. Businesses should maintain detailed records of human creative involvement at every stage of the AI-assisted production process. This includes records of prompt design, editorial review, selection choices, and any modifications made to AI output before it is used. These records serve two purposes. They support the argument that a human author was involved – reinforcing the tool-analogy approach – and they provide the evidentiary foundation for a registration application. Courts applying the tool-analogy doctrine will need to see evidence of human creative choices. Documentation creates that evidence.
The second priority is contractual structure. Employment agreements, contractor agreements, and technology licensing contracts should be reviewed and updated to address AI-generated output explicitly. Assignment clauses should clearly cover output produced with AI assistance. Where the company is the downstream user of a third-party AI tool, the terms of service and licensing agreement with the tool provider should be reviewed for IP ownership provisions. Some providers claim rights over outputs. Others disclaim all rights. The contractual position directly affects the company's ability to assert ownership.
The third priority is registration where possible. Even under conditions of uncertainty, registration with the National Copyright Directorate creates a date-stamped record. If a dispute arises later, that record is valuable. The registration should accurately describe the nature of the work. Where AI involvement is significant, the application should be prepared carefully to present the human contributions in the most legally coherent way. Misrepresenting the nature of authorship in a registration carries its own risks.
The fourth priority is monitoring legislative and judicial developments. Argentina's legislature has begun receiving proposals for AI regulation, and the IP question is likely to be addressed as part of a broader AI regulatory regime. The trajectory of regional legislation in Brazil and the international discourse around AI Act compliance in Europe will influence the Argentine debate. Businesses with significant AI-generated IP assets should engage in that debate – through industry associations or directly – to ensure that whatever legislative solution emerges reflects commercial realities.
A non-obvious risk in this area concerns training data. Many Argentine businesses deploying AI tools have not audited the training data underlying the models they use. If the model was trained on copyrighted material without licence, outputs generated by that model may carry infringement risk. Under Argentine civil liability doctrine, the deploying company could face claims from rights-holders whose material was used in training. This risk is distinct from the authorship question but is practically connected to it. A comprehensive IP strategy for AI-generated works must address both the ownership of outputs and the provenance of inputs.
The economics of inaction are significant. A business that does not structure its AI-generated IP position now may find, when it seeks to enforce rights or execute a licensing transaction. That its assets are either unprotected or encumbered by unresolved ownership questions. The opportunity cost of that outcome – lost licensing revenue, failed enforcement, reduced asset value in a sale or investment transaction – is material. The steps to mitigate that risk are available today and are not disproportionately costly relative to the commercial value at stake.
Outlook: the legislative horizon and what practitioners are watching
Argentina's IP system will need to respond to AI-generated works. The question is not whether reform will come, but when and in what form. Several indicators are worth monitoring.
The first is the trajectory of Argentine AI regulation more broadly. Several legislative proposals are circulating that address algorithmic accountability, transparency obligations for AI systems, and – in at least one proposal – questions of IP ownership for AI-generated content. None has yet been adopted. The pace of Argentine legislative reform on technology matters has historically been slower than commercial adoption. However, the political salience of AI is rising, and the legislature is under increasing pressure from industry groups and civil society to act.
The second indicator is regional convergence. If Brazil enacts an AI regulation that includes IP provisions, and if Mexico follows a similar path, the pressure on Argentina to harmonise will increase. The Mercosur framework, while not a vehicle for IP legislation in the way that EU law operates in Europe, does create commercial incentives for regulatory alignment among member states. Argentine practitioners are watching Brazilian legislative developments closely for this reason.
The third indicator is the IP office's administrative practice. The National Copyright Directorate's handling of registration applications that disclose AI involvement will shape practical outcomes before any court or legislature acts. If the Directorate develops a published policy – accepting AI-assisted works under conditions, or refusing works where human creative input is below a threshold – that policy will function as quasi-regulation. Industry engagement with the Directorate during any consultation process is therefore a practical advocacy opportunity.
The fourth indicator is court practice. As AI-generated IP disputes reach Argentine courts – and they will, given the volume of AI-generated content now in commercial circulation – judicial reasoning will accumulate. The first significant decision from the National Civil Court of Appeals or the Supreme Court will be highly consequential. Businesses with pending disputes touching on AI-generated works should consider the litigation strategy implications carefully. The first major case shapes the doctrine for those that follow.
Practitioners in Argentina also note the influence of the academic and scholarly community. Argentine IP doctrine has historically been shaped by academic commentary in ways that are less typical in common law systems. The body of legal scholarship on AI-generated works is growing. That scholarship will inform both judicial reasoning and legislative drafting. Businesses with a significant stake in the outcome have an interest in supporting rigorous legal analysis in this area.
Frequently asked questions
Q: Can an AI system hold intellectual property rights in Argentina?
A: No. Under Argentina's intellectual property legislation, only natural persons can be recognised as authors. An AI system lacks legal personality and cannot hold or exercise IP rights. Ownership of AI-generated output therefore vests in a human or legal entity connected to the creation process, subject to how that connection is structured.
Q: How long does it take to register an AI-assisted work in Argentina, and what does it cost?
A: Registration with the relevant IP authority in Argentina typically takes several weeks to a few months, depending on the type of work and the completeness of the application. Government fees are modest and vary by category. Legal fees for preparing the filing and advising on authorship strategy add to the total cost. Early registration is strongly recommended, as it creates a date-stamped record that is valuable in any subsequent dispute.
Q: Is it a misconception that Argentina's copyright law automatically protects AI-generated output?
A: Yes, that is a common misconception. Automatic protection under Argentina's copyright regime applies to original works of human authorship. Output produced autonomously by an AI tool without identifiable human creative choices falls outside the traditional scope of protection. Businesses that assume AI-generated content is protected without further analysis risk losing exclusivity over commercially valuable material.
About Ferraz & Whitmore
Ferraz & Whitmore is an international law firm based in Lisbon, advising business clients across 46 jurisdictions. Our team combines Portuguese civil law expertise with English common law tradition to deliver cross-border legal solutions in AI and technology law, intellectual property protection, and technology licensing. We work with technology companies, institutional investors, and in-house legal teams operating across the Americas who need results-oriented counsel on AI-generated works, software liability, algorithmic accountability, and cross-border IP strategy. Our Americas practice covers Argentina, Brazil, Mexico, Colombia, Chile, and neighbouring markets, with direct experience advising on digital services regulatory questions and AI Act compliance for companies with transatlantic operations. Engaging a lawyer in Argentina with cross-border IP and technology experience is essential when AI-generated assets form part of a business's commercial value. As an international law firm serving Argentina and the wider region, Ferraz & Whitmore helps clients build defensible IP positions from the ground up. To discuss your AI-generated works strategy in Argentina or across the Americas, contact us at info@ferrazwhitmore.com.
Disclaimer: This publication is provided for informational purposes only and does not constitute legal advice. The information herein should not be relied upon as a substitute for professional legal counsel tailored to your specific circumstances. Ferraz & Whitmore assumes no liability for actions taken or not taken based on the contents of this material. For advice regarding your particular situation, please contact info@ferrazwhitmore.com.