The Bard's Dilemma: Preserving Human Story in the Age of Machine Narration

A Literary Essay

I. The Mad King’s Gift

In the seventh century, an Irish king named Suibhne Geilt went mad at the Battle of Mag Rath. According to the medieval text Buile Shuibhne, he sprouted feathers, flew into the treetops, and spent the rest of his life wandering Ireland as a wild prophet—composing some of the most haunting poetry in the Irish tradition while living backward through time, seeing the future as memory and the past as prophecy.

I am his descendant, thirty-seven generations removed.

I tell you this not to romanticize madness or claim prophetic insight, but because Suibhne’s story—part king, part bird, part bard—captures something essential about the creative act itself: the necessity of dwelling in multiple modes of being simultaneously. The writer must be part technician and part mystic, part architect and part dreamer, grounded in craft while reaching toward something that transcends it.

Today, as I watch artificial intelligence systems generate unlimited volumes of plausible fiction with a few keystrokes, I find myself thinking about Suibhne in his tree, looking backward and forward at once. We stand at a similar precipice—a moment when the very nature of storytelling is being transformed, when the boundary between human and machine creation threatens to dissolve entirely.

The question is not whether AI will generate stories. It already does, prolifically and convincingly. The question is whether we can preserve something essential about human creativity in the process—something that goes missing when we reduce storytelling to pattern-matching and statistical prediction.

This essay is about that preservation, and about what we might recover in the attempt.

II. The Content Deluge

Walk into any digital bookstore and you’ll find them: novels with generic covers, plausible premises, competent prose. Read a chapter or two and something feels hollow—not poorly written exactly, but absent of that quality that makes literature matter. The metaphors are apt but predictable. The characters move through emotional arcs that feel algorithmically determined. Everything is correct in the way that machine translation is correct: conveying information without resonance.

Some of these books are fully AI-generated, though they rarely advertise this fact. Others are human-outlined, machine-drafted, human-edited—a hybrid that makes attribution nearly impossible. The market floods with this content because the economics are irresistible: why pay a writer for three years of work when you can prompt a language model for three hours?

We are witnessing the industrialization of storytelling.

This is not new, of course. The pulp magazines of the 1920s and 30s commodified fiction. Romance publishers in the 1970s created assembly-line formulas. What’s different now is the scale and the invisibility. AI-generated content is unlimited, cheap, and increasingly indistinguishable from human work. It doesn’t tire, doesn’t need health insurance, doesn’t suffer writer’s block.

The economic threat to human writers is real and immediate. But there’s a deeper, more existential problem: when we flood the cultural ecosystem with machine-generated text trained on human literature, we create a feedback loop that gradually degrades the very thing the machines learned from. Each generation of AI content becomes training data for the next, diluting the distinctiveness, the strangeness, the essential humanness that makes literature matter.

We risk creating a culture of plausible surfaces—a million novels that read like novels without being novels in any meaningful sense.

III. The Authorship Question

What does it mean to “write” a story when AI generates the prose?

Current discourse offers two unsatisfying answers:

The Rejectionist Position: “AI has no place in creative work. Use of AI invalidates authorship entirely.”

This position preserves purity but abandons powerful tools. It’s the literary equivalent of insisting that writers use quills and parchment because typewriters are mechanical assistance. Technology has always mediated creative work. The question isn’t whether to use tools, but how to use them without losing ourselves in the process.

The Techno-Optimist Position: “AI is just another tool. The human who prompts it is the author.”

This position embraces capability but evades the hard question: if I type “Write me a story about a dragon” and ChatGPT generates 5,000 words, am I really the author? I chose the premise, yes. But did I make the thousand small decisions that constitute the actual work of writing—the rhythm of a sentence, the choice to reveal character through gesture rather than exposition, the deliberate ambiguity in a key scene?

Both positions miss something essential: Authorship is not about who holds the keyboard. It’s about who makes the decisions that carry meaning.

When a human writer crafts a sentence, they’re not just selecting words—they’re making choices about emphasis, rhythm, implication, resonance. They’re deciding what to reveal and what to withhold, which metaphor serves the theme and which merely decorates. These decisions emerge from intentionality, from the writer’s lived experience and moral imagination, from their attempt to say something true about being human.

Can an AI make such decisions? In a sense, yes—it selects words based on patterns in its training data. But these are not meaningful decisions. The AI has no lived experience, no moral imagination, no truth it’s trying to articulate. It arranges plausible patterns without understanding why those patterns matter.

The crisis, then, is this: As AI becomes capable of generating more sophisticated prose, we risk losing the distinction between human meaning-making and machine pattern-matching.

IV. The Bardic Tradition

To understand what we’re losing, it helps to look back at what we’ve already lost.

Before the printing press, before literacy became universal, stories were participatory events. The bard didn’t simply recite a fixed text to passive listeners. The audience shaped the telling through their reactions—when they leaned forward, when they laughed, when they fell silent. The bard adapted in real-time, extending scenes that resonated, abbreviating ones that didn’t, improvising details based on who sat around the fire that night.

This wasn’t lack of authorial control. It was a different kind of authorship—collaborative, responsive, alive. The story existed in the space between teller and listener, shaped by both, owned by neither alone.

The shift to print brought tremendous gains—preservation, distribution, individual artistic vision. But we lost something too: the participatory nature of storytelling, the sense that stories were events we created together rather than objects we consumed alone.

Interactive fiction has tried to recover this, with limited success. Choose-your-own-adventure books offered branching paths but felt mechanical. Video games created participatory narratives but often sacrificed literary quality. Tabletop role-playing games came closest—a human game master facilitating collaborative storytelling—but required multiple people synchronizing schedules, and the quality varied wildly with the game master’s skill.

What if we could recover that bardic tradition in a new form? What if technology could enable genuinely participatory storytelling that preserves literary quality—not by replacing the human game master, but by amplifying their capability?

This is where AI could actually serve human creativity rather than supplant it.

V. The Technical Possibility

I am, by profession, a software architect specializing in autonomous systems—specifically, aviation systems where the stakes for human safety are absolute. In that domain, we’ve learned a critical principle: automation must augment human judgment, never replace it.

When designing autopilot systems for drones, we don’t build black boxes that make autonomous decisions. We build systems with clear authority boundaries: the machine can execute complex maneuvers, but the human pilot retains ultimate command. The pilot can always intervene, always understands what the system is doing, and bears moral responsibility for outcomes.

This principle—which I call the Agency Paradox—applies equally to creative systems:

Machines may execute, optimize, and render—but must never originate human meaning, moral responsibility, or intentional direction.

In practical terms for storytelling, this means:

AI can generate prose drafts within human-defined constraints
AI can offer multiple options for human evaluation
AI can maintain continuity and check consistency
But AI cannot decide character motivations
AI cannot resolve moral dilemmas
AI cannot choose plot directions
AI cannot introduce irreversible narrative events

Every decision that carries narrative meaning must pass through an explicit human approval point—what I call a decision gate. These gates are unavoidable, clearly presented, and recorded in a provenance ledger that proves human authorship.

The result is a system where humans originate meaning and AI renders it into prose. The human authors the story; the AI articulates it. The distinction is preserved, traceable, verifiable.

VI. ChronicleEngine: A Demonstration

Let me make this concrete.

I’m building a narrative system called ChronicleEngine—not as a commercial product (though it may become one), but as a proof of concept for these principles. Think of it as an AI-augmented game master for solo interactive fiction, but with ironclad guarantees about human authorship.

Here’s how it works:

You load a story module—think of it as a world with defined rules, themes, and constraints, created by a module designer (the first tier of authorship). You create a character and declare their goal. The system presents you with narrative context and offers choices, but you—the player (the second tier of authorship)—make every meaningful decision.

Want your character to seek revenge rather than redemption? That’s a character goal decision that requires your explicit choice from available options. The system presents alternatives, shows consequences, and blocks until you decide. It cannot default, cannot suggest a “best” option, cannot proceed without your conscious choice.

When you make this decision, the system records it in a provenance ledger: what you decided, what alternatives existed, when you decided, why (if you choose to explain). This ledger is your proof of authorship.

The AI then renders your decision into prose—describes your character’s burning desire for vengeance, generates dialogue that reflects their motivation, narrates events consistent with their goal. But it’s rendering your meaning, not originating its own.

At every major decision point—plot direction, moral dilemmas, irreversible events—you pass through another decision gate. The cumulative effect is a story that’s demonstrably yours: every meaningful choice traced back to your human decision, with the AI serving as a sophisticated rendering engine.

The system maintains complete event history. You can replay your story and it unfolds identically—same states, same outcomes, provably your story. You can export the provenance ledger to show exactly where human agency operated and where AI rendering occurred.

This isn’t theoretical. I’m implementing it using a deterministic edge AI architecture that keeps your story data on your device (privacy preserved), operates fully offline, and costs a fraction of cloud-only systems.

VII. Three Tiers of Authorship

ChronicleEngine embodies a three-tier model of distributed authorship:

Tier 1: Framework Creator (my role, initially) I define what constitutes a “meaningful decision,” how decision gates function, what constraints the system enforces. I create the architecture that guarantees human agency. This is authorship at the level of possibility—determining what kinds of stories can be told.

Tier 2: Module Creator (community world-builders) Module creators define specific story worlds—their themes, rules, available character types, narrative primitives. They decide “in this world, revenge is possible but redemption is not,” or “magic has these costs and these limits.” This is authorship at the level of constraint—shaping the boundaries within which stories unfold.

Tier 3: Player (end users) Players make decisions within module constraints—goals, choices, moral positions, plot commitments. They create the specific narrative through their decisions. This is authorship at the level of instantiation—bringing particular stories into being.

All three tiers are essential. All three require humans. The AI serves all three but replaces none.

This model addresses a deeper question: in an age of AI, what counts as “authorship”? Not typing words—that’s execution. Not having ideas—that’s inspiration. Authorship is making the decisions that generate meaning under constraint.

When those decisions are made by humans, recorded as such, and traceable through provenance, we preserve authorship even when machines handle the prose.

VIII. What We Recover

If we can build systems that preserve human agency while leveraging AI capability, what do we recover?

The Bardic Tradition, Technologically Mediated

Interactive fiction becomes truly participatory again—not passive consumption, but co-creation with clear authorship boundaries. The “bard” (the system) facilitates, but the audience (the player) shapes the story through real decisions.

Literary Quality in Interactive Form

Because AI handles prose generation within human-defined constraints, interactive fiction can achieve literary sophistication previously impossible in branching narratives. Every path doesn’t need to be written in advance—the system generates it, but from your meaningful decisions.

Verified Human Creativity

In an age when anyone can generate thousands of pages of plausible fiction overnight, provenance becomes valuable. “This story was created with ChronicleEngine” means “this story has verifiable human agency at every meaningful decision point.”

Publishers might eventually require such verification. Readers might demand it. “Human-authored” becomes a certification, not just a claim.

A Framework for Other Creative Domains

The principles extend beyond interactive fiction. Screenwriting tools, novel-drafting assistants, poetry generators—any creative AI system could implement decision gates and provenance to preserve human authorship while leveraging machine capability.

We could establish industry standards for guaranteed human authorship, preventing the content dilution crisis before it fully metastasizes.

IX. The Writer-Engineer-Bard

I began this essay with Suibhne Geilt—the mad king who became a poet, dwelling in trees, seeing backward and forward simultaneously. I return to him now because his story illuminates my own strange synthesis.

I spent a decade in aviation software, building autopilot systems for autonomous drones. I became fluent in the language of safety-critical systems, deterministic control, human-in-command architectures. I learned how to make machines powerful without making them dangerous.

But I’ve also been writing—working on a fantasy trilogy that explores the origin story of the wise mentor archetype through a grimdark king who becomes a backward-living prophet. (Yes, it’s inspired by Suibhne. How could it not be?)

For years, these felt like separate identities: the engineer and the writer. One technical, one creative. One concerned with determinism and safety, the other with ambiguity and beauty.

ChronicleEngine emerged when I realized they’re not separate at all. The engineering discipline I learned building autonomous systems—preserving human agency, maintaining provenance, enforcing boundaries—applies directly to the crisis facing writers in the AI age.

The technical and the creative converge in a question both domains must answer: How do we keep humans in command when machines become capable?

In aviation, we answer with pilot authority and safety protocols. In literature, we answer with decision gates and authorship verification. Different implementations, same principle.

And beneath both lies something older: the bardic tradition, the human need to tell stories together, to create shared meaning through narrative. Technology should serve that tradition, not replace it.

I am, it turns out, what my ancestors would recognize: a bard who uses different tools but serves the same purpose—facilitating stories that matter because humans chose them.

X. The Stakes

Let me be clear about what’s at stake.

If we allow AI to generate stories autonomously, with humans as mere editors or prompters, we will:

1. Devalue Human Creativity Economically Why pay writers when machines produce unlimited content for pennies? The profession of author becomes unsustainable for all but the most successful.

2. Erode Literature Culturally As AI-generated content floods the market, trained on human literature but devoid of human intentionality, the signal-to-noise ratio collapses. Finding genuinely human work becomes archeology.

3. Lose Something Essential Philosophically Stories matter because they’re acts of meaning-making by conscious beings who understand mortality, love, loss, joy—who live the human condition they’re depicting. When machines generate plausible narratives without understanding, without living, we get surfaces without depth.

But if we build systems that preserve human agency while leveraging AI capability, we can:

1. Amplify Human Creativity Writers gain powerful tools that handle execution without abdicating authorship. Interactive fiction becomes viable as serious literature. New forms become possible.

2. Establish Standards Provenance systems create verifiable human authorship, enabling readers to distinguish genuine human work from machine approximation.

3. Recover Lost Traditions The participatory nature of oral storytelling returns in a new form—technologically mediated but humanly authored.

This isn’t Luddism. It’s not rejecting technology. It’s insisting that technology serve human flourishing rather than replace it.

XI. A Call to Fellow Travelers

I write this essay knowing that many readers will be skeptical.

Writers will suspect I’m building tools that enable their obsolescence while claiming to preserve their agency.

Technologists will think I’m overcomplicating what should be a simple prompt-to-output pipeline.

Literary critics will question whether interactive fiction can ever achieve true literary merit.

And publishers will see only the economic headache of verifying authorship in an age when verification is hard.

To all of you, I say: The crisis is real. The dilution is happening. We need solutions, not just critique.

ChronicleEngine is my attempt—one architect’s effort to apply hard-won lessons from safety-critical systems to the preservation of human creativity. It may not be the only solution. It may not even be the best solution. But it’s a serious effort to address a serious problem.

I invite collaboration:

Writers: Help me understand what authorship means to you. What decisions feel essentially human? What could AI handle without stealing your voice?

Technologists: Challenge my architectures. Propose better ways to enforce decision gates. Find the holes in my provenance system.

Literary Critics: Hold this vision accountable to standards of serious literature. Don’t let me settle for gamified choice trees when what we need is genuine narrative art.

Publishers: Consider what industry standards for human authorship might look like. What would you need to see to certify a work as genuinely human-authored?

Philosophers and Theologians: Help me think clearly about agency, intentionality, and meaning. This is ultimately a question about what makes us human.

The code I’m writing will be open source. The philosophy I’m developing will be public. I’m not trying to own this solution—I’m trying to start a conversation that leads to better solutions.

XII. Conclusion: The Bard in the Tree

Suibhne Geilt spent his final years in the trees, composing poetry that would survive him by more than a millennium. The medieval scribes who recorded his story couldn’t decide if he was mad or prophetic—perhaps both, perhaps neither.

What matters is that his words survived because they were his—a human being’s attempt to articulate the strange experience of living, told in a voice that could only be Suibhne’s. The kings who commanded armies are forgotten. The bard in the tree is remembered.

We stand at a moment when machines can generate unlimited plausible words, but they cannot give us Suibhne’s voice. They cannot give us any human voice, because they have no lives to speak from.

The question before us is whether we’ll preserve the distinction that makes literature matter—the presence of a human consciousness trying to articulate truth through story—or whether we’ll accept a flood of eloquent surfaces that say nothing because no one meant them.

I believe we can preserve human creativity while leveraging AI’s capability. I believe we can recover the participatory nature of storytelling while maintaining literary quality. I believe we can build systems that prove authorship even when machines handle execution.

But I don’t believe any of this will happen automatically. It will require intention, effort, and collaboration. It will require writer-engineer-bards who can bridge the technical and the creative. It will require all of us deciding that human meaning-making matters enough to preserve.

Thirty-seven generations removed from a mad king in a tree, I’m attempting my small part: building a system that lets humans tell stories together with AI as our rendering engine, not our replacement. Proving that technology can amplify human creativity without dissolving it.

The bard is still in the tree, looking backward and forward, seeing what was and what might be. The question is whether we’ll listen—and whether we’ll choose to remain in the conversation.

Author’s Note:

This essay describes work in active development. ChronicleEngine is not yet publicly available, but the philosophy and architecture are being developed openly. The code will be released as open source when sufficiently mature.

I welcome correspondence from writers, technologists, publishers, and anyone interested in preserving human creativity in the AI age.

Stephen Sweeney
Principal Architect, Autonomous Systems
Writer-Engineer-Bard
Descendant of Suibhne Geilt

Contact: stephen@agentincommand.ai
Project: github.com/stephen-sweeney/ChronicleEngine (forthcoming)
Blog: agentincommand.ai

Acknowledgments:

To the medieval scribes who preserved Buile Shuibhne, ensuring Suibhne’s voice reached us across centuries.

To J.R.R. Tolkien, who showed that serious literature could be epic fantasy, and that sub-creation is a profound human act.

To every writer fighting to preserve human creativity in an age of machine generation.

And to my ancestors, who kept the bardic tradition alive long enough for me to attempt its recovery in this strange new form.

This essay is available under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Share freely, adapt as needed, but preserve attribution and the core commitment to human agency.