Imagine a world where your deepest, unarticulated desires—those you are not even consciously aware of—are brought to life. Now, consider if simply articulating those desires in natural language is all it takes to make it happen. While this may seem like an exaggeration, it is increasingly becoming a reality, especially in the realms of language and image generation with the aid of artificial intelligence. Today, we encounter a myriad of AI-generated images varying vastly in quality. Why do some captivate us instantly while others fail to make an impression? The answer lies in how well these images resonate with our internal aesthetic desires.
[1]The secret to creating impactful AI-generated art involves communicating precisely what we seek—whether it be in the style of Van Gogh or Monet. It requires specifying details such as colors, hues, spatial relationships, sensations and more. The deeper our understanding of the styles and our ability to articulate detailed desires, the more accurately AI can meet our expectations. This brings us to recognize the unsung heroes: the individuals crafting the backend prompts that guide AI. They meticulously articulate, in great detail, what many of us cannot even fully perceive ourselves. With the increasing prevalence of such sophisticated AI-generated artworks, which require a specialized skill set to create, a critical question arises: Does the prompt writer, this unsung hero, deserve any intellectual property protection?
Copyright Protection of Prompts
The answer seems straightforward, doesn’t it? Simply grant copyright protection to the prompts.[2] But is this really the right approach? Let’s delve into its implications. Imagine a scenario where a particularly useful prompt, popular with many, is protected by copyright. What does enforcement look like in this case? Consider a person using this copyrighted prompt on ChatGPT to generate compelling AI content. How do you enforce the copyright? How would you even become aware of the infringement?
Unlike publicly sold online content, where the copyright owner can monitor usage, the utilization of prompts is predominantly in private scenarios, making detection extremely challenging. This scenario illustrates that the rights granted to a prompt writer may, in practice, be hollow and ineffective. Is this fair, or even what they truly need?
Furthermore, this raises side effects and policy issues. For example, should the Large Language Model (LLM) companies be liable for allowing copyrighted prompts to be run on their platforms? Internet companies faced similar challenges and are now largely exempt due to the requirement for a Digital Millennium Copyright Act (DMCA) takedown notice before a lawsuit can proceed. See 17 U.S.C. § 512. This DMCA exemption is practical because copyright owners can identify infringing activities when the infringing works are publicly accessible. However, for prompts, except for the platforms themselves, it is nearly impossible for a prompt writer to determine who has used their copyrighted work.
If LLM platforms were obliged to monitor such use, it would significantly disrupt the user experience. Note that the threshold for copyright is lower than for patents; it generally requires originality, not novelty. Thus, with a surge of copyrighted prompts flooding an LLM platform, the platform finds itself in a dilemma: balancing the need to maintain a seamless user experience, which allows users to input any prompt, against the substantial costs of enforcing copyright claims from prompt writers. Moreover, which users would be comfortable using a platform that constantly monitors them for potential copyright infringements? What happens if a user innocently inputs something that happens to be protected by copyright? Would the platform then report her?
If platforms do not take this responsibility, another likely option left for prompt writers to enforce their rights is to speculate based on the content generated by users. Yet, most often, content created from prompts is used privately and remains undiscovered. Even if used publicly, how can one ascertain that it was produced from a specific copyrighted prompt, given the inherent randomness in how LLMs generate responses? Even for identical prompts, the output can vary. Under Federal Rules of Civil Procedure Rule 11, plaintiffs need a reasonable basis to file a lawsuit, but given the variability of LLM outputs, can they reasonably claim that the content was generated from their copyrighted prompt?
Given the above challenges, it is likely impractical to use copyright to protect prompts for all parties involved—the prompt writers, the LLM platforms and the users. Ultimately, this approach could render the rights of prompt writers ineffectual, or worse, ignite unnecessary disputes driven by a sense of helplessness and the need to speculate.
Copyright Protection of AI-Generated Contents
Current law makes it very difficult, if not impossible, to copyright AI-generated content.[3] Yet, paradoxically, the law might be turning things upside down by considering copyright for prompts while denying it for the content AI produces. [4]
The underlying principle of the law is that the work must have “human authorship.”[5] However, arguably, all man-made creations inherently involve human creativity. For instance, since LLMs cannot generate output without prompts written by humans, doesn’t this satisfy the legal requirement for human authorship? This leads to the next question: what exactly constitutes “human authorship”? The Copyright Registration Guidance’s answer is: “users do not exercise ultimate creative control over how such systems interpret prompts and generate material. Instead, these prompts function more like instructions to a commissioned artist—they identify what the prompter wishes to have depicted, but the machine determines how those instructions are implemented in its output.”[6] However, even in the case of a commissioned artist, the law regarding works made for hire would transfer authorship to the entity that commissioned the work. See 17 U.S.C. § 201(b). In other words, drawing an analogy between AI and a commissioned artist suggests that authorship would more naturally belong to the entity that commissioned the work: here, the prompt writers.
Alternatively, if one views AI as merely a tool, then there is even less justification for denying copyright on AI-generated content. Indeed, when someone uses a pen to write a poem, no one questions the authorship of the person holding the pen. Why, then, do some argue that when a prompt is inputted and the machine produces an output, the prompt writer should be denied authorship? Many explanations exist, but one might be our unfamiliarity with AI’s capabilities. We see a significant transformation from a simple input to a complex output and feel disconnected from the authorship because the process is beyond our understanding. However, by denying our authorship in this transformation, we are essentially dismissing the influence of our own desires in shaping the AI’s output. Our desires are the very power that can guide the AI to fulfill our deepest needs. We can cultivate this power by recognizing and rewarding the content generated by prompt writers, who articulate our desires at a conscious level.
Granting copyright protection to AI-generated content in exchange for the disclosure of prompts, while establishing that these prompts are not protected by copyright when used with LLMs, could resolve many issues. First, it offers incentives for prompt writers to share their prompts with the public, enhancing our awareness of our own thoughts and desires. Second, users would be free to use the prompts however they wish without restriction. Third, it simplifies oversight for the LLM platforms. Instead of filtering prompts, platforms need only ensure that the outputs are not identical to the copyrighted content initially generated by the copyright owner.
Indeed, this approach mirrors the patent system, where a limited monopoly is granted for a useful machine or method in exchange for disclosing how to make them. This has significantly advanced technology over the past 200 years. Similarly, we could advance human understanding by rewarding individuals’ ability to articulate the intricate details of human desires, thus revealing what we truly want, and helping us either achieve these desires or let them go when we understand their true nature. Whether this is worth pursuing is a decision for Congress.
While concerns might arise that using prompts to generate content is so easy that it could overwhelm the Copyright Office, this issue isn’t new; people are already free to register copyrights for countless photographs, yet only do so when it makes economic sense due to application fees. To prevent abuse, the Copyright Office could set an appropriate fee for registering AI-generated content and might require the applicant to disclaim all training materials used to develop the LLM. These measures would help balance economic effects and achieve equilibrium.
Therefore, it seems we might be overly cautious about awarding copyright to AI-generated content for reasons we don’t fully understand. Upon considering practical factors, it might make sense to rethink our approach. Are we indeed doing things upside down?
***************************
In the end, let’s return to our opening scenario: Imagine a world where your deepest, unarticulated desires—those you are not even consciously aware of—are brought vividly to life. Now, if simply articulating those desires in natural language could make it happen, what price tag would you place on this magic?
[1] This article aims to address the question from a policy perspective, given the current uncertainty in the law and regulations. Rather than detailing what the law currently states, this article focuses on the potential policies driving ongoing legislation and rulemaking.
[2] See Application Process for Registration of Works with Artificial Intelligence-Generated Content （June 28, 2023） at 11 (“And I guess I would just add to that, you know, if the prompts contained enough text or whatever material, you know, we would certainly evaluate it and make our decision. But it would be important to remember that if we did register the prompts, all we would be doing is registering the text or whatever content that's been submitted to us”).
[3] Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence (March 16, 2023) at 4 (“If a work’s traditional elements of authorship were produced by a machine, the work lacks human authorship and the Office will not register it. For example, when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology—not the human user.”)
[4] Id. n.27 (“While some prompts may be sufficiently creative to be protected by copyright, that does not mean that material generated from a copyrightable prompt is itself copyrightable”).
[5] Id. at 2.
[6] Id. at 4.