generative AI Archives – Andrew Gibson Author Narrator Editor Curator

Introduction: The Double-Edged Sword of Generative AI

The last few years have seen artificial intelligence leap from research labs into everyday life. Tools that can generate images, compose music, write essays, and even narrate audiobooks are no longer speculative novelties—they’re mainstream. As generative AI becomes faster, cheaper, and more accessible, it’s tempting to see it as a revolutionary force that will boost productivity and unlock new forms of creativity. But beneath the surface of this techno-optimism lies an uncomfortable truth: much of this innovation is built on the uncredited labour of human creators. AI does not invent from nothing; it remixes the work of writers, musicians, and artists who came before it. If these creators can no longer sustain their livelihoods, the very source material that AI depends upon could vanish.

AI Doesn’t Create—It Consumes and Repackages

At its core, generative AI is a machine of imitation. It ingests vast amounts of text, audio, or visual data—almost always produced by human beings—and uses statistical models to generate plausible imitations of that content. While it may seem impressive that an AI can write a poem or narrate a story in a soothing voice, it’s critical to understand where that ability comes from. These systems are trained on real works created by real people, often scraped from the web without consent or compensation. The machine doesn’t understand the meaning of its output; it only knows what patterns tend to follow other patterns. When creators can no longer afford to produce the original works that fuel these systems, the well of quality data will inevitably run dry.

The Hollowing Out of Voice Work and Storytelling

Few sectors have felt the AI crunch more viscerally than the world of audiobook narration. Platforms like ACX, once bustling with human narrators offering rich, emotionally nuanced performances, are increasingly confronted by the spectre of synthetic voices. These AI narrators are trained to mimic tone, pacing, and inflection—but what they deliver is, at best, a facsimile. They lack the lived experience, instinct, and intuition that make a story come alive. Narration is more than enunciation; it’s performance, interpretation, and empathy. By replacing voice artists with digital clones, platforms risk reducing literature to something flavourless and sterile—a commodity stripped of its soul.

Software Developers: Collaborators or Obsolete?

The anxiety isn’t limited to creative fields. Developers, too, are questioning their place in an AI-saturated future. With tools like GitHub Copilot and ChatGPT able to generate code in seconds, it’s fair to ask whether programming is becoming a commodity task. But while AI can write code, it cannot originate vision. Consider EZC, a project built using AI-assisted coding. The AI wrote lines of JavaScript, yes—but the concept, purpose, and user experience all stemmed from a human mind. Writing code is only a fraction of what development truly entails. Problem definition, audience empathy, interface design, iteration—all these remain stubbornly human.

Should We Use AI to Replace What Humans Do Best?

There’s a compelling argument for using AI in domains that defy human capability: mapping the human genome, analysing protein folds, simulating weather systems. These are tasks where data volume, speed, and pattern recognition outstrip our natural capacities. But the push to replace things humans do best—like storytelling, journalism, art—is not progress. It’s regression masquerading as innovation. AI thrives on what already exists, but it doesn’t dream, it doesn’t reflect, and it certainly doesn’t feel. Replacing human creativity with predictive models creates a feedback loop of derivative content. Over time, the result isn’t abundance—it’s entropy.

Swarm AI and the Illusion of Independence

Some argue that AI’s future isn’t as a tool but as a fully autonomous agent. Imagine swarms of AI agents identifying market needs, writing business plans, building applications, and launching them—without human input. Technologically, this may be within reach. Ethically and existentially, it’s a minefield. Even the most sophisticated AI lacks the moral compass and cultural context that guide human decision-making. Left unchecked, these systems could flood the world with unoriginal, unvetted, and even harmful content. The question isn’t whether AI can act independently, but whether it should—and who decides the guardrails.

Co-Creation, Not Replacement: A Path Forward

There’s a more hopeful vision of the future: one in which AI is a powerful collaborator, not a competitor. In this model, humans provide the spark—an idea, a question, a vision—and AI accelerates the execution. The most impactful work comes from this synergy: where human insight shapes the direction and AI helps scale it. Instead of replacing narrators, we could use AI to offer alternative formats, translations, or accessibility features. Instead of replacing developers, we could use AI to automate routine tasks, freeing up time for higher-level design thinking. It’s not a matter of resisting AI—but insisting it be used ethically, responsibly, and in service of human creativity, not as a substitute for it.

Conclusion: Don’t Let the Well Run Dry

AI has extraordinary potential—but without a steady stream of human imagination to draw from, that potential is finite. We must resist the temptation to replace human creators simply because it’s cheaper or more scalable. What makes art, software, journalism, and storytelling valuable is the messy, intuitive, and lived experience behind them. If we hollow out the professions that produce meaning, we risk filling the world with noise. This is not about anti-AI paranoia—it’s about pro-human stewardship. The future of creativity doesn’t belong to machines; it belongs to the people bold enough to use machines as tools, not replacements.

Press Play to listen to this Article about: Intellectual Property and AI

The Fragile Foundations of Intellectual Property in a Post-Human World

Intellectual Property (IP) law is predicated on a simple, almost quaint notion: that creativity originates from a human mind. For centuries, this idea formed the bedrock of legal systems that sought to reward originality, incentivize innovation, and protect creators from exploitation. Copyrights, trademarks, and patents all assumed a world where authorship could be attributed, originality could be proven, and infringements could be identified and punished. Artificial Intelligence, however, has no interest in playing by these rules. It creates not from intention but from interpolation, not from inspiration but from ingestion. The moment we allowed machines to mimic human output, we introduced a crisis that the old IP framework is wholly unequipped to handle.

AI Generates, But Who Owns the Output?

When a generative AI produces a novel, a painting, or even a working piece of software, the immediate question becomes: who owns it? Is it the person who typed in the prompt? The team that trained the model? The company that owns the servers? Or is it no one at all? The law currently has no satisfactory answer, and that legal void is being filled — not with regulation — but with millions of new AI-generated artifacts flooding the internet daily. This isn’t a legal grey area anymore; it’s a full-blown epistemological collapse. We no longer know where content comes from, let alone who should be credited or paid for it.

Fair Use Was Never Meant for This

The companies behind the largest AI models argue that their training data falls under “fair use.” This is a legal doctrine designed to allow commentary, parody, and critique — not industrial-scale ingestion of copyrighted material to produce infinite derivative content. Every time a model generates something that sounds like Taylor Swift, reads like Margaret Atwood, or paints like Greg Rutkowski, it does so based on absorbed data. If the model never “sees” these creators’ work, it cannot emulate them. But if it does see them, and profits are made without consent or compensation, how is this anything but theft in slow motion? Courts are starting to weigh in, but existing law was never built to arbitrate between authors and algorithms. We’re asking a Victorian legal structure to moderate a space-age dispute.

Enforcement Is Impossible at Scale

Let’s say IP rights do technically survive. Let’s say the courts rule that training on copyrighted work without permission is illegal. Let’s even say watermarking AI output becomes mandatory. None of that will matter. AI tools are proliferating at such speed and volume that enforcement becomes nothing more than whack-a-mole with a blindfold. How do you pursue legal action against a user in an uncooperative jurisdiction using an open-source AI model trained on pirated datasets to generate content that may or may not resemble your work? The burden of proof is on the creator, the costs are prohibitive, and the damage — once done — is irreparable. Enforcement, in this new era, is like chasing ghosts with a broom.

IP Assumes Scarcity — AI Offers Infinity

At the heart of IP law lies the assumption that creative works are finite and special. A song, a novel, a design — each is protected because it represents time, effort, and unique human insight. But AI erases that scarcity. Once a model is trained, it can generate an infinite supply of anything, in any style, at any time. This not only devalues individual works but also reduces the incentive to create them in the first place. Why buy a stock photo, commission a design, or license music when a comparable substitute can be generated for free? The market is shifting from one of scarcity to one of surplus, and IP law cannot function in a world where the marginal cost of creation is zero.

The Disintegration of Attribution and Provenance

Provenance — the history and authorship of a creative work — used to matter. It was how collectors valued art, how scholars verified texts, and how courts resolved disputes. But in the age of AI, provenance is rapidly becoming irrelevant. Most AI-generated content lacks metadata that can trace it back to a clear source, and even when watermarks are added, they’re easily stripped or bypassed. Worse, many AI models now run locally or in decentralized environments, completely beyond the reach of regulatory oversight. The result is a digital Wild West where no one knows what’s real, who made it, or who should be held accountable. In this landscape, attribution becomes a nostalgic ideal — not a practical tool.

The Economic Impact on Human Creators

The collapse of enforceable IP rights has immediate consequences for anyone who creates for a living. Writers, artists, musicians, filmmakers, and developers are watching as their work becomes raw material for systems that can replicate it, remix it, and render it obsolete. As AI-generated content floods the internet, the market value of human-made work is driven down. Platforms and clients increasingly seek quantity over quality, speed over skill, and price over provenance. Some creators will adapt, of course — becoming prompt engineers, curators, or performance-based brands. But many will not. For them, the age of AI isn’t a new opportunity; it’s an extinction event.

Legacy IP Models Are Dead Weight in a Fluid Ecosystem

Large content platforms — YouTube, Spotify, Amazon — rely on rigid, centralized IP systems. But AI-generated content doesn’t fit cleanly into that infrastructure. It’s too fast, too amorphous, and too anonymous. These platforms will either have to overhaul their systems to support new forms of authorship or accept that a growing percentage of their content cannot be reliably traced or monetized under old models. Startups and decentralised platforms, meanwhile, are embracing the chaos. They’re not asking who owns the content; they’re asking how to scale it, optimize it, and sell it. And they’re winning. The more flexible the platform, the less IP matters.

A Glimpse at What Comes Next

So if traditional IP dies, what replaces it? The most likely answer is reputation-based economies, where success depends less on what you create and more on who you are. Creators will trade in trust, visibility, and community — offering experiences, interactions, and ongoing value rather than isolated products. Watermarking and provenance systems, possibly based on blockchain or other decentralised ledgers, may help retain some sense of authorship, but they will be voluntary, not enforceable. Licensing may evolve into subscription-style access to models, templates, and toolkits rather than individual pieces of media. But the idea of “owning” a melody, a sentence, or a visual style? That’s going away. Forever.

Conclusion: Intellectual Property Isn’t Evolving — It’s Disintegrating

AI doesn’t respect Intellectual Property, not because it’s malicious, but because it operates on principles entirely alien to human creativity. It doesn’t ask permission, cite sources, or respect boundaries — it just generates. And once content becomes infinite, attribution becomes irrelevant, enforcement becomes impractical, and ownership becomes obsolete. In such a world, clinging to old legal frameworks is like trying to copyright the wind. The sooner we accept that, the sooner we can start building new models that reflect the strange, synthetic creativity of this emerging era. IP isn’t being disrupted. It’s being obliterated.

Tag: generative AI

Human Creativity in the Age of AI: Innovation or Erosion?