Three classics of science fiction, rebuilt as games and simulations with AI.
The three games presented below represent how an AI system interpreted classic stories. They are not designed to replace the readings for students. Rather, they offer a creative possibility to see how a new technology like generative AI might complement or complicate our relationship with literature.
Le Guin's narrator builds a city of perfect happiness before your eyes, then shows you its foundation: a single child, locked in a cellar, whose misery is the price of everyone else's joy. Every citizen knows. Most make peace with it. A few walk out of the city and do not come back. At four thousand words, it remains one of the most efficient moral traps in the language — and the narrator keeps asking the reader's permission to spring it.
Omelas — The Walk drops you inside the Festival of Summer as a first-person visitor. Six voices of the narrator drift through the city as points of golden light; you find them, listen, and answer the questions the story asks of you. Before entering, you record a prediction about what a "perfect" city costs. At the end, the game hands you a reflection report with your name on it — and your earlier answer waiting beside what you now know.
Assign the story first; the game's questions fall flat without it. The prediction-versus-reflection report is a ready-made writing prompt for first-year composition or an ethics seminar. And since Le Guin's narrator keeps asking the reader for permission, discuss what changes when the asking becomes literal.
On a world lit by six suns, true darkness comes once every 2,049 years — and with it the Stars, a sight no mind on the planet has ever survived intact. Asimov's scientists calculate the eclipse, predict the madness, and prepare anyway, while the archaeologists point to layer after layer of civilizations that burned themselves down reaching for light. It is the genre's great parable about the limits of knowing something you have never experienced.
Rather than retell the plot, NIGHTFALL — Kalgash hands you the premise as a survival problem: gather wood while the suns still shine, plant torches, manage a sanity meter that drains in darkness, and reach the Observatory before red Dovim sets. Two endings — The Stars and Dawn — depending on whether you kept a flame between yourself and the sky.
This one works as a systems model: students experience scarcity, preparation, and panic rather than reading about them. Afterward, ask what the simulation leaves out — the scientists, the cultists, the journalist — and why those omissions matter. The gap between the model and the text is the discussion.
Ellison's AM — a military supercomputer that destroyed humanity — keeps the last five people alive to torment for eternity, because hatred is the only thing it has left. The narrator's single act of mercy, and what it costs him, gives the story its title. It is the genre's bleakest answer to the question of machine intelligence, and nearly sixty years on, still one of its most assigned.
NO MOUTH is a first-person descent: two levels, eight chambers, a flashlight, and dialogue written as original tribute rather than quotation. It is explicitly framed as a fan-made, non-commercial homage that points players back to the source — its own opening credits insist that you read the story it honors.
Best suited to AI-ethics and digital-culture courses. Start with the irony — an AI built this homage to fiction's most malevolent AI — and let students argue whether that is tribute or trespass. Preview it before assigning; like the story, it is deliberately upsetting, and an opt-out alternative should be offered.
Disclosure is pedagogy. If we ask students to be transparent about how they use AI, the materials we hand them should model the same standard.
Three stories were chosen because they are widely anthologized, commonly taught, and each takes a different stance toward technology and the reader.
Each game was created in conversation with Claude (Anthropic): described, generated, playtested, and revised across multiple sessions by one instructor. No code was written by hand.
Claims made by AI were verified by a human being and edited for clarity.
This page itself was designed and drafted with AI assistance, then edited and approved by its human author. That is the standard we ask of students, so it is the standard here.
The three games linked from this page — and the page itself — were created with generative AI (Claude, by Anthropic), directed, playtested, and edited by Marc Watkins. "The Ones Who Walk Away from Omelas" © Ursula K. Le Guin; "Nightfall" © Isaac Asimov; "I Have No Mouth, and I Must Scream" © Harlan Ellison. All rights remain with the authors and their estates. The games are non-commercial educational tributes and reproduce no story text. None of this replaces the reading.