At first glance, this legal phrasing appears to target a specific technology. But when parsed with analytical neutrality, it reveals a profound legislative double standard. The law explicitly punishes the human creator for the harms of synthetic media, yet steadfastly refuses to grant them credit for its creative triumphs. This hypocrisy exposes a fundamental flaw in our legal and philosophical definitions of agency, labor, and authorship.
By deconstructing the relationship between the human prompter and the machine, we can uncover a compelling truth: the person who directs generative AI is performing the exact same function as a high-level executive. If we hold them responsible for the risks, we must grant them the rewards.
The Executive Analogy: The President and the Administration
To understand why the current legal status quo is failing, we must look at how human organizations operate. Consider a nation’s President and their administrative officials.
When a President wants to enact a policy or draft a historic address, they do not write every word or calculate every economic variable themselves. They provide high-level intent, strategic frameworks, and rigid boundaries. The administrative officials—highly trained agents—take those instructions and execute them.
In doing so, these officials use their own brains. They make millions of autonomous micro-decisions, choose specific vocabularies, and format the final output. Yet, under the legal doctrine of political accountability and organizational management, the credit and the blame belong entirely to the President. The officials do not own a copyright over the methodology of their execution; their labor belongs to the enterprise directed by the executive.
Generative AI operates on identical merits. A user prompting an AI music generator or an image model is not merely pushing a button; they are acting as the Executive Producer. The AI serves as the digital administration. The user dictates the genre, the emotional resonance, the structural boundaries, and the thematic constraints. The machine then processes those inputs and executes the micro-details—the exact chord progressions, the micro-rhythms, or the pixel values.
If the functional input (executive command) and the functional output (completed work) are identical in both scenarios, then the President and the AI prompter are doing the exact same job.
The Fallacy of the "Black Box"
The primary counter-argument used by legal essentialists is the concept of the AI's "black box." Critics argue that because an AI’s decisions are based on probabilistic mathematics rather than conscious choice, the human user cannot predict or control the exact expression of the output, thus invalidating their claim to authorship.
But this argument crumbles under close scrutiny. Human administrative officials also possess a black box: the human subconscious.
When a human official executes a command, their output is subtly altered by biological randomness and systemic bias. If they are fatigued, they might word a sentence differently. If they carry personal biases from their upbringing, it will warp the tone of their work. They draw from a lifetime of unpredictable emotions, suffering, and joy.
Generative AI simulates these exact human fluctuations, simply substituting biology with math. Through hyper-parameters like "Temperature" or "Top-P," developers introduce deliberate randomness into an AI model, forcing it to make unpredictable, "creative" choices. Furthermore, the AI inherits the collective biases, cultural nuances, and emotional structures embedded within its billions of data points of human training text.
When a user prompts an AI with words like "melancholy jazz," the machine calculates the statistical probability of what humans perceive as sadness. Whether a subtle shift in tone comes from a tired speechwriter or an AI operating at a high temperature setting, the final page looks the same. To the observer, the biological black box and the mathematical black box are functionally indistinguishable.
The Gymnastics of the Prompt
Furthermore, the legal status quo severely underestimates the labor involved in modern prompting. There is a persistent myth that AI creation is effortless. In reality, advanced AI users engage in intense mental "gymnastics" to achieve a precise vision.
They spend days crafting "perfect executive prompts." This is a rigorous process of iterative editing, testing parameters, adjusting weights, and constant course correction. The prompter must continuously reject the machine’s autonomous deviations and herd the algorithm back toward the desired framework—much like a director commanding a crew or a leader managing a stubborn bureaucracy. This high-level orchestration is a massive investment of human intellect, time, and intent. It is the very definition of creative labor.
The Legal Splitting of Blame and Credit
Despite these parallel merits, global courts and regulatory bodies have split the concept of control into a contradictory double standard:
• For Liability (The Blame): The law completely accepts the executive analogy. Under frameworks like India’s IT Rules or Western AI safety mandates, if a user prompts an AI to generate a defamatory song, a deepfake scam, or non-consensual intimate imagery, the user is held fully liable. The law rightly rejects the defense that "the AI's black box decided the exact words, not me." The human set the chain of events in motion; therefore, the human bears the blame.
• For Copyright (The Credit): The moment the same user asks for ownership over a beautiful, legally compliant AI creation, the law reverses its logic. Global precedents state that because a machine executed the final expression, the work lacks "human authorship" and belongs to the public domain.
This creates a glaring paradox: If a human is powerful enough to control an AI to the point of being criminally punished for its harms, how can they be viewed as too powerless to be credited for its creativity?
The Judicial Status Quo: Simple vs. Executive Prompts
This deadlock has pushed global judicial bodies into a state of chaotic fragmentation as they attempt to define the exact line between a simple request and true executive authorship.
In the United States, the legal system has drawn an aggressive line against machine autonomy. In landmark decisions like Thaler v. Perlmutter, federal courts affirmed that copyright requires a biological human baseline, completely shutting down claims where the AI is named as the sole author. When human creators tried to claim ownership over AI-assisted works by showcasing their intricate prompt workflows—such as Kristina Kashtanova’s comic book Zarya of the Dawn—the U.S. Copyright Office fractured the protection. They granted copyright to the human-written text and the overall compilation layout, but completely stripped protection from the individual images generated by Midjourney, declaring that a user does not exercise sufficient predictive control over the machine’s ultimate visual output.
Parallel battles are playing out in Indian courts, where judges are grappling with the interpretation of Section 2(d) of the Copyright Act, 1957, which defines an author in relation to computer-generated works as "the person who causes the work to be created." The definition seems to open a clear door for the executive prompter. However, the practical application remains highly unstable. The Indian Copyright Office famously granted co-authorship status to an AI painting assistant named the RAGHAV Intellectual Property Rights Cell alongside its human owner, only to later issue a notice to withdraw it upon realizing the implications. These global flip-flops show that the judiciary is desperately measuring physical tinkering and manual adjustment rather than recognizing the systemic value of executive intent.
The Legislative Blueprint: Algorithmic Work-for-Hire
To resolve this legal paralysis, we do not need to reinvent the legal wheel; we need to adapt an existing mechanism: the Work-for-Hire doctrine.
In traditional corporate law, specifically under Section 201(b) of the U.S. Copyright Act and similar commercial structures in India, if an employee creates a software program or drafts a corporate report, the employee does not own the copyright. The law employs a legal fiction that vests the initial authorship directly in the employer or the entity that commissioned the work. The employer provides the framework, the compensation, and the executive mandate, while the employee provides the mechanical execution.
We must codify an "Algorithmic Work-for-Hire" framework into modern digital law. Under this legislative blueprint, the statute would explicitly declare that an autonomous algorithm cannot hold property or be recognized as a legal person, but its mechanical outputs are legally bound to the entity that issued the prompt.
This mechanism creates complete harmony between liability and credit. By treating the AI prompter as the legal employer of the machine, the user automatically receives 100% of the copyright protections and commercial rights. Simultaneously, it legally locks them into 100% of the civil and criminal liability if that same machine generates unlawful synthetic content. The "President" gets the credit because they legally own the apparatus of execution, eliminating the middle ground where AI-generated content floats in a lawless public domain.
The Technical Limits of Provenance: The Tuition Problem
Even if the law adopts the Algorithmic Work-for-Hire framework to reward the prompter, it faces a massive underlying economic crisis: the training data dilemma. If the AI "administration" learned how to write music or paint images by analyzing millions of copyrighted works by real human artists without compensation, the prompter is essentially benefiting from stolen education. The AI creator owes "tuition fees" to the teachers.
Technologists have long argued that we can solve this by tracking data provenance—using digital watermarks and cryptographic tracking (such as the C2PA standard) to calculate micro-royalties for the original artists whenever an AI generates content. However, this solution hits brutal technical walls.
First, metadata and digital watermarking are highly fragile. A digital watermark embedded into an AI image or audio track can be easily stripped away by taking a screenshot, running the file through basic compression algorithms, or re-recording the audio. Once the metadata is severed, the link to the original training source is lost entirely.
Second, deep neural networks operate through non-linear mathematical compression. When an AI model trains on a million songs, it does not store copy-pasted fragments of those songs. Instead, it extracts abstract mathematical weights and distributes them across billions of parameters. When a user generates a new jazz song, the AI draws fractions of percentages of influence from thousands of different artists simultaneously.
Calculating the exact contribution of a specific human artist to a single generated note is a computational nightmare. True micro-attribution requires tracing back through billions of active neural pathways in reverse—a process so resource-intensive that it breaks the scalability of the technology. Until computer science creates untamperable, low-latency tracking systems, the law cannot accurately calculate the tuition fees owed to human artists, leaving the distribution of creative credit fundamentally messy.
Finding the Truth
To reconcile this tension, society must abandon outdated, romanticized definitions of authorship that require physical execution. We must look at the situation with absolute functional neutrality.
The current legal refusal to grant copyright to AI prompters is not based on a lack of logic—it is an artificial barrier maintained by governments to protect human markets from being overwhelmed by machine-speed generation. While the anxieties regarding artistic displacement and monopolization are valid, they must be solved through targeted anti-monopoly regulations and data-licensing laws, not by fracturing the logical consistency of authorship and liability.
Technology has forced us to confront the fact that executive intent is creative labor. If the law is brave enough to hold the executive responsible when the machine sins, it must be just enough to grant them the copyright when the machine sings.
- inputs by Team Aatmeeyataa Patrekaa
Comments