The Holodeck, the Canvas, and the Algorithm: Navigating the Legal Paradox of Generative AI

Disclaimer: The following is not legal advice. It is simply a random civilian philosophizing about the intersection of technology, law, and human nature.

In the 24th century, aboard the USS Enterprise-D, Lieutenant Reginald Barclay had a problem. Brilliant but crippled by social anxiety, the engineer retreated to the ship’s holodeck to run highly realistic, intimate, and entirely unapproved simulations involving holographic replicas of his crewmates — most notably, Counselor Troi and Dr. Crusher.

When the senior staff discovered this in the Star Trek: The Next Generation episode “Hollow Pursuits,” it presented a fascinating dilemma. By any modern standard, discovering a colleague has created a hyper-realistic virtual sandbox to act out inappropriate fantasies involving your exact likeness is a profound violation of boundaries. Yet, the approach of Captain Picard and the crew was a masterclass in theoretical, idealized HR: they aimed to help rather than accuse and punish. They recognized that while his actions were intensely uncomfortable for the subjects, Barclay was dealing with his own psychological struggles in a simulated, private environment. No physical harm was inflicted.

This futuristic scenario forces us to look backward at the very bedrock of our justice system, which relies on two absolute pillars: “innocent until proven guilty” and “the burden of proof is with the accuser.” We judge tangible actions, not internal thoughts.

From the Mind to the Canvas

Taking the Barclay precedent back to the present day, the baseline is clear. If an individual is fantasizing about another person in the privacy of their own mind, it is not a crime. It is unprovable beyond a reasonable doubt, and the legal system does not, and cannot, police thoughts.

But what happens when that internal fantasy crosses the threshold into the physical world? Suppose this individual, lacking a 24th-century holodeck but possessing a physical canvas and a paintbrush, paints an indecent portrait of the person they are fixated on.

What is the legal precedence here? Can the (potential) victims patent or copyright their face to prevent this?

The short answer is no. You cannot copyright your own physical “look.” Copyright law protects original works of authorship fixed in a tangible medium. Ironically, the painter owns the copyright to the painting, even if the subject is someone else.

However, the victim has other legal avenues to stop the perpetrator:

• The Right of Publicity: This prevents the unauthorized commercial exploitation of an individual’s likeness. If the painter attempts to sell, publish, or exhibit the work, the victim can intervene.
• Defamation and False Light: If the painting damages the victim’s reputation or falsely implies they posed for the indecent work, civil lawsuits apply.
• Intentional Infliction of Emotional Distress (IIED): If the act of creating and sharing the painting is deemed extreme, outrageous, and designed to cause severe psychological trauma.

The Deepfake Escalation

Technology, however, has dramatically lowered the barrier to entry. We have moved far beyond the physical canvas. With the advent of Photoshop, and ultimately deepfakes and advanced Generative AI, anyone can create hyper-realistic, non-consensual imagery in seconds.

Society is currently scrambling to establish legal protections against this. Historically, we relied on the civil torts mentioned above. Recently, jurisdictions have aggressively updated Non-Consensual Intimate Imagery (NCII) laws — often called “revenge porn” laws — to explicitly include synthetic media and deepfakes. Yet, enforcement remains incredibly difficult. When the perpetrator is an anonymous user operating behind a VPN, the burden of proof becomes an insurmountable wall for the accuser.

The Liability Paradox: Why GenAI is Treated Differently

This brings us to a glaring hypocrisy in the tech industry. For decades, Adobe has provided Photoshop to the masses. It is a tool that has undoubtedly been used to forge documents, create defamatory images, and produce harmful content. Yet, Adobe is not routinely sued for the crimes of its users. The consensus has always been that the software is merely a tool; the liability rests entirely with the user wielding it.

Why, then, are modern AI software developers facing massive legal scrutiny, with society demanding they take responsibility for the artifacts their users create?

The shift in blame stems from a misunderstanding of system architecture. When viewing generative AI as the central processing unit of a new computational era, the user’s prompt acts as the program, and the RAG-based knowledge base serves as the memory. Critics argue that because the “CPU” was trained on vast amounts of human data, it is no longer just a passive brush — it is an active participant in the creation. But fundamentally, if the AI only executes the parameters defined by the human “programmer,” shouldn’t it still be protected under the same tool-liability frameworks as Photoshop?

The Copyright Reversal: A Thought Experiment

Let’s push this legal philosophy to its absolute limit. If society ultimately forces software developers to take responsibility and legal liability for the artifacts their users generate, shouldn’t those developers also take the credit — and the intellectual property rights?

Currently, AI-generated art cannot be copyrighted because it lacks “human authorship” and intent. But what if AI companies restructure their operations?

Instead of offering a “software service,” they could operate as an “art agency.” The workflow shifts slightly: a user submits a prompt, which legally acts as a “commission request.” An employee at the tech company receives these requests in a batch system, clicks a button to acknowledge the workflow, and the AI tools are deployed to produce the requested painting.

In this scenario, human intent is re-established. The tech company has the intent to fulfill a client’s commission and uses its proprietary systems to execute the art. If the developer is forced to bear the legal risk of the creation, it stands to reason they should simply claim the copyright of the resulting artifact, treating users merely as clients who purchased a localized license.

Final Thoughts

We are standing at a crossroads. We cannot treat Generative AI as an autonomous agent when it comes time to assign legal blame, but dismiss it as a lifeless, un-copyrightable algorithm when assigning intellectual property rights.

Whether we are dealing with a canvas, a server farm, or a holodeck, our legal frameworks must decide what these systems actually are. Until we define the boundary between the tool, the creator, and the liability, we will remain stuck in a paradox, waiting for the legal safety protocols to engage.

For — as they stand today — the Federation, the owner of the USS Enterprise, on which the Holodeck is located and provided, will get sued for Lieutenant Barclay’s holo-fantasies (and possibly get sued for copyright infringement Professor Moriarty it produced as well).