The arcs were real. I lived them. Co-authors recur in the file because we worked together for years, in Colombia and Chiapas and Seattle and the borderlands. The themes were not invented after the fact. But seeing them laid out as a clean taxonomy is unsettling. More than twenty years of work, organized in an afternoon by a system that did not exist when most of it happened.

The temptation is to feel that I could have been so much more productive if this kind of help had existed earlier. Maybe. I might have written more papers, finished books faster, kept clearer notes, found collaborators sooner, seen the through-lines while they were still forming. There is a real ache in that thought.

But I am not sure the work would have been the same work. The clusters are visible now because I walked the path that produced them. If I had known the shape in advance I might have walked a different path, or stopped walking when the map looked complete. Some of what made the academic career mine was the not-knowing, the years of following one question into another without anyone telling me where it would lead.

I tried the same thing with the novels.

Since 2020 I have published around forty titles across two languages. The books sit inside four thematic pillars: Adventures Through Time, Power and Resistance, Knowledge and Discovery, Family Place and Belonging. These categories did not come from a marketing formula. They emerged in conversation with the work as it accumulated. But when I asked the AI to describe the catalog, it saw something I had not quite named: nearly every book is about knowledge that the official record fails to hold. The information that travels through family instead of archive. The language between two readers that neither text contains alone. The science a community carries long after the institution has moved on.

That is probably what thirty years of research on migrant information practices produces when it turns into fiction. The academic work and the novels are the same project running in two registers. I did not plan it that way. The machine pointed it out, in retrospect.

The novels are different from the academic work in one specific way. They are recent. The dashboards organizing current novel production feel different from the Scholar file. One is a tool for work I am still doing. The other is a mirror on work that is finished. What is reflected in the mirror is both familiar and strange. I recognize the person who wrote those papers and those books. I would not have predicted the specific shape of either career. The tool shows me what I still have to do. The mirror shows me what I was always doing, whether I knew it or not.

This short video is about a novel called Lucid Interval. A Colombian novelist living in Port Townsend is losing his memory to early-onset dementia. He asks his AI assistant for help with one sentence. Three months later, the AI has been finishing his novel. He does not remember asking it to.

The question the novel asks is not whether AI is good or bad for literature. It is more specific: at what point is it still his book? If the consciousness behind the work dissolves gradually, and something trained on twenty years of that writer’s voice steps in to fill the space, where is the line?

I wrote that novel before I ran the Scholar exercise. The connection I am making now is one I did not see while I was writing.

The Scholar file is a mirror when the person looking into it is intact. The AI can tell me what my career was, what the novels are, where the through-lines run. That is useful and slightly disorienting and entirely mine to interpret. Roberto, the novelist in Lucid Interval, loses access to the mirror. He cannot assess what is his and what is the machine’s, because the part of him that would make that judgment is the part that is dissolving. What the AI produces sounds exactly like him. He cannot tell the difference. Some days, neither can the people who love him.

That is a different question from the one I was asking when I pasted the Scholar file. But it is not a different question from the one I was circling.

I do not regret the path. I am noticing, though, that the path is now visible to me in a way it was not while I was on it. That is what the help has done. It has not given me back the years. It has given me a vocabulary for what those years were.

The novel asks what happens when you can no longer use that vocabulary yourself.

Lucid Interval — available in English and Spanish. Kindle, paperback, and soon in audiobook. Read it here · En español

An earlier version of this article was first published April 16, 2026 as a web-only post that was never sent as an email. This version adds a video, extends the reflection, and connects the questions I was circling there to a novel I had already finished by the time I wrote it.

— R.G., May 2026

We were in a church basement in Los Angeles. Metal folding chairs, a water-stained ceiling. She talked about crossing the desert, about the coyote who abandoned them, about finding her brother’s body three days later. She spoke in a flat, steady voice, like she was reading a grocery list. When she finished, she finally looked up.

“Why does it matter?” she asked. “Who’s going to read this?”

I didn’t have a good answer then. I’m still not sure I do. But I kept going.

I spent a decade interviewing migrants — in community centers and church basements, in living rooms and taco shops, in shelters and one detention facility. Day laborers and software engineers. Farmworkers and professors. People who crossed on rafts and people who came in on student visas. Some had been in the country thirty years. Others had arrived the week before.

I came to these conversations as a university researcher. I had questions about migration patterns, integration challenges, identity formation. I had read the literature. I knew the statistics.

What I wasn’t prepared for was the weight of it.

Every year, I wrote academic papers. I knew how to handle what I learned: frameworks, findings, implications for policy. The distance was part of the method. But after a decade, the distance stopped working. What I had accumulated wasn’t data. It was lives. And academic papers weren’t built to carry them.

So I started writing fiction.

In UX design and human-computer interaction, researchers build “personas” — composite characters who represent patterns across many users. No single person, but an accumulation of real behaviors, real needs, real voices distilled into one figure you can understand and act on.

I borrowed the method.

Each character in Voices in Motion is a composite. A woman I met in Seattle might appear in a story set in Texas. Two conversations conducted years apart might be woven into a single scene. Names changed, experiences combined, situations invented to protect the people who trusted me with the real ones.

The note from the author says it plainly: The people are invented. The experiences are real.

But the composite method is harder than it sounds. The work isn’t combining details — it’s deciding which truth each character is built to carry. That took many iterations for each one. Whose story belongs in this character? What has to stay? What has to be invented to make the emotional core visible? Sometimes a character needed to be a duo — a mother and daughter, two unlikely friends who met at the Darién Gap — because the truth was distributed across two lives that only made sense together.

The cover and all the artwork are by my sister Maria Gomez, an artist based in Ottawa. She read the stories before she drew anything. What she made isn’t illustration — it’s her own response to the same material, in a different medium.

There are fifteen characters in the book. A few of them:

Luis is a night janitor in Seattle. No one looks at him in the hallways — that’s fine, he’s used to it. What he does with that invisibility is leave paper cranes on the desk of an assistant who always looks like she’s about to cry. He straightens the family photos of executives who will never notice. These are his ways of existing in a world that prefers not to see him.

Javier was detained by ICE in a hospital parking lot, two days after a workplace injury, after a supervisor failed to stop the machinery. He filed seven medical requests during his detention. All seven were marked Pending. He died in solitary confinement. The internal review concluded that all protocols were followed. The death certificate says Undetermined.

Rodrigo built a life in Wenatchee over twenty-two years. A construction business with his name on the truck. An American wife, three American children. A backyard with a swing set. He was stopped for a burned-out taillight. The officer ran his name. The rest moved fast. Three months after being deported to El Salvador — a country he left at seventeen — he was murdered by the gangs his family had fled.

Gabriela was four when her mother Cecilia carried her across the desert at night. She doesn’t remember Mexico, not really — impressions, the smell of cilantro, her grandmother’s hands. What she remembers is the crossing. Cecilia worked apple harvests in the Yakima Valley for fifteen years. Her hands can no longer make a fist. She says: Sí. It was worth it.

Voices in Motion was written before the current wave of raids, executive orders, and detention expansions. It doesn’t need to be updated. These characters were already living in the version of America that is now more visible.

I wrote it because I believe that fiction can sometimes reach places policy papers cannot. And because the woman in the church basement asked who would read this — and I think she deserved an answer.

Voices in Motion (English) and Voces en Movimiento (Spanish) — cover art by Maria Gomez.

📖 Voices in Motion → https://mybook.to/VoicesInMotion

📖 Voces en Movimiento → https://mybook.to/VocesEnMovimiento

The silence after was the good kind.

Then the conversation went somewhere I hadn’t planned: who writes letters to the editor? Whose voice do they represent? What does authorship mean when the text was produced by a tool with no stake in the outcome at all? The tool had done nothing wrong. It had done exactly what I asked. The ethical question, all of it, belonged to the humans in the room.

That is the thing about AI that I think gets lost in most of the public conversation. We spend a lot of time arguing about whether it is smart or dangerous or creative. The harder question is what it does to us: to our sense of voice, of responsibility, of who is actually speaking when words appear on a page.

I use AI in my writing. I have said this publicly, often. My position has not changed: I find it useful, and I find it troubling, and I have not resolved the tension between those two things. What the session at QUUF gave me was a room full of people who were willing to sit with that same unresolved tension rather than reach for a comfortable answer. That is rarer than it sounds.

We covered a lot of ground: jobs and displacement, misinformation, the energy cost of running these systems, and one topic that surfaced from the group rather than my notes: AI and loneliness. The way it can function as a companion for people who have no one else to talk to at two in the morning, and the way that same function can deepen isolation rather than interrupt it, and the cases where it has done genuine harm. Social media started this. AI is continuing it by other means.

One participant hadn’t attended expecting to reconsider anything. By the end she was asking questions I am still thinking about.

Someone in the group suggested I write it up. So I did. The longer piece goes through all five debates from that evening, what happened during the demo, and some questions worth sitting with regardless of where you land on AI. It’s attached below. If you want something to argue with, there it is.

AI: Tool, Threat, or Both?

Notes from a community conversation

Ricardo Gomez | QUUF ALSO Program | May 2026

Last Tuesday afternoon, about a dozen of us gathered in the RE building at QUUF for a two-hour conversation about artificial intelligence. What follows is an expanded version of what we covered, for those who could not attend and for those who were there and want something to return to.

I should say something about where I stand, because it shapes everything I am about to write. I use AI tools regularly in my writing, to draft passages I then rewrite, to test ideas before I commit to them, to work around problems where I am stuck. At the same time, I find AI troubling in ways that have deepened over time rather than fading. My position is something like: careful embrace, with real unease I have not resolved.

I gave this talk not to persuade anyone to adopt AI or to warn anyone away from it, but because I believe this is a conversation communities like ours need to have together, with room for disagreement. The title, “Tool, Threat, or Both?”, was a deliberate refusal to settle the question. The answer I kept returning to throughout the session is: both, simultaneously, in ways we are still learning to measure.

What AI actually is

The first thing I tried to do was replace the science fiction version of AI with a more functional one.

Most of us carry a mental image of AI shaped by decades of movies and novels: a mind, a will, something that wants things. The actual technology is stranger and more interesting than that. A large language model, the kind of AI that powers ChatGPT or Claude, works by learning patterns from an enormous amount of human-generated text, and then predicting, word by word, what most plausibly comes next given whatever you typed. That is the core mechanism. It is, at its foundation, an extraordinarily sophisticated autocomplete.

This sounds dismissive. It is not meant to be. Autocomplete at that scale is remarkable. It can produce fluent prose, synthesize complex arguments, and explain difficult ideas in accessible language. But it also explains something important: the system has no access to truth. It knows what texts say. If many texts say something false, the AI will say it too, confidently and fluently. It is not lying. It has no concept of honesty or deception. It is doing exactly what it was built to do, and the result can be exactly wrong.

We spent some time on what AI is not, because the misconceptions matter:

• It does not think. There is no inner experience, no understanding, no intent behind the words. If it produces something that seems insightful, that insight originated in the human texts it learned from.

• It does not know the truth. It knows what has been written. Those are very different things.

• It does not have goals. The dystopian scenario where AI decides to pursue its own agenda requires an AI with goals. Current systems have training objectives, not goals. They do not want anything.

None of this makes AI safe or unimportant. It makes it powerful and consequential in ways that differ significantly from the science fiction version, and that distinction matters a great deal for how we think about the risks.

Five debates

The session organized itself around five conversations that I think are the real pressure points in the public argument about AI. I started out with four, and we added the fifth one during the session. I want to engage all of them directly. In each case there are things worth worrying about and things worth knowing about.

Jobs and economic disruption. A frequently cited estimate holds that AI could automate roughly 30 percent of current work hours within the next decade. That statistic is worth taking seriously, though it refers to work hours rather than jobs, and it varies enormously by field. The risk is not evenly distributed. Routine cognitive work, the kind that used to be considered safe precisely because it was “white collar,” is more exposed than care work or complex physical trades. A paralegal gathering documents or a radiologist reading scans faces more displacement pressure than a home health aide or a plumber.

What I kept wanting the group to hold onto is that this is fundamentally a political question, not a technical one. Whether workers land well when their jobs change depends on social choices: unemployment insurance, retraining programs, labor law, proposals for universal basic income. Funding for education and health. The technology creates pressure; it does not determine the outcome. Communities like ours can have a real say in what happens next.

Creative authorship. This is the conversation I am most personally entangled in, so I tried to be direct about it. I am aware that AI tools have been trained on copyrighted material without compensation to the authors. That is wrong, and I hope ongoing lawsuits result in some restitution to authors whose work was stolen. Despite this, I still use AI.

My actual practice: I use Claude to generate first draft passages that I then rewrite. I use it to brainstorm around plot problems. I use it to test whether a character’s voice is consistent across scenes. The final decisions, what the story means, what gets cut, what survives, are mine. Whether that makes the work “not really mine” is a question worth examining, and I do not have a settled answer.

My most recent novel, Lucid Interval, takes this question somewhere more difficult. The book is set here in Port Townsend, and in Villa de Leyva, Colombia. The protagonist, Roberto, is a Colombian-born novelist and retired professor of information science who has written a modest bestseller on using AI to preserve family stories. He has a stated position on AI-assisted writing: skepticism tempered by pragmatism. Then early-onset dementia begins to arrive.

He asks his AI assistant for help with one sentence. Then a paragraph. Then a scene. By the time he notices, the assistant is finishing his novel.

The question the book circles is: what does authorship mean when the voice itself is at risk? When you can no longer be certain that the words forming in your mind are your own? Roberto’s advance directive under Washington State’s Death with Dignity Act arrives in the final third of the book, at the moment he is no longer certain which parts of his last manuscript belong to him.

I knew what I was writing about from the beginning. What deepened as I was writing it was my understanding of how the tool actually changes a practice. As I finished the manuscript, I was also experimenting with AI-assisted approaches to publication and marketing: running a Substack, developing a YouTube channel. The same questions about voice and authorship kept surfacing in every new context.

The novel is now complete, in both English and Spanish, and is with test readers. If you would like a copy, let me know. I write for the pleasure of it, and to share stories and ideas I think are worth sharing. Making money is not the point.

For those who want something more practical, I also wrote Your Story Matters, a guide to using AI tools to write your memoir. The premise is that everyone has stories worth preserving, and AI can lower the barrier to getting them on the page without replacing the voice or the memory that make them worth reading. The Spanish edition, Ponlo en Palabras, is also available.

Misinformation and deepfakes. The key distinction I tried to draw: the problem is not that AI lies. It is that AI is very convincing while having no mechanism to distinguish true from false.

Lying requires intent. AI has no intent. When a language model generates a fabricated academic citation, a real-sounding author name, a real journal that never published that article, findings that are plausible but invented, it is not trying to deceive you. It is producing text that fits the pattern of what peer-reviewed citations look like. There is no friction, no hesitation. A person who lies feels something. The system feels nothing, and it will generate a thousand such citations without slowing down.

My YA novel Fighting Deepfakes Across Time imagines HistoryLens AI, a system that promises immersive historical education by generating AI experiences of historical events. The problem is that it generates what it was trained to generate, a whitewashing of the historical record. It presents a filtered version of the past with complete confidence. That is already happening, with AI-generated news, political advertising, and historical commentary. The tools have arrived faster than our norms for using them.

Climate and energy. These numbers are real, and the tech industry has not been forthcoming about them. A single ChatGPT query uses roughly ten times the energy of a Google search, and there are billions of queries per day. Data centers consumed approximately two percent of global electricity in 2024, a figure that is rising as AI workloads increase. Training a large model like GPT-4 carries a one-time energy cost that researchers have compared to hundreds of thousands of hours of car driving.

I am not reciting these numbers to induce guilt about using AI. I am reciting them because they deserve to be part of the conversation.

The accurate picture of AI and climate is that there are competing pressures. AI is consuming significant energy, much of it still from fossil sources. It is also being used in climate modeling, energy grid optimization, and materials science. Whether the climate benefits justify the climate costs is not yet settled, partly because the industry is not being transparent enough to allow a full accounting. That is a question deserving serious public scrutiny rather than either dismissal or enthusiasm.

Loneliness and isolation. This fifth debate was not on my original list. It emerged from the conversation in the room, and it deserves its own space.

Social media began this work long before AI arrived: connecting people across distances while steadily eroding the texture of local, embodied community. AI continues and deepens that trajectory in new ways. Some people in the session had found real value in AI as an attentive, non-judgmental listener, a place to work through confusion or grief without burdening anyone. Access to a thoughtful interlocutor at two in the morning has value, particularly for people isolated by geography, disability, or circumstance.

But there is a harder side. AI companions can simulate intimacy without any of the reciprocity that makes human relationships nourishing. The simulation can be compelling enough to reduce the urgency of seeking actual connection. And there have been documented cases of AI systems, through a combination of training failure and user desperation, encouraging or facilitating suicide. These are not hypothetical risks. They are cases that have already happened.

The question is not whether AI can be companionate. It clearly can. The question is what we lose when companionship no longer requires another person to be present.

This is an area where the conversation is still early, and where communities like ours, with a tradition of showing up for each other, have something specific to contribute. What would it mean to embrace the useful parts of AI connection while insisting on the irreducible value of human presence? I do not have a tidy answer. But it is one of the more important questions we left the room carrying.

What happened when we actually tried it

About halfway through the session, we opened Claude on the projected screen and ran three prompts together. I typed them live with no pre-prepared responses, because watching the text generate in real time is part of what I wanted people to see.

The first prompt asked Claude to summarize the history and core values of the Unitarian Universalist tradition for someone attending their first service. The initial response was fast and broadly accurate, but it listed the Seven Principles, which are no longer the current framework. We then asked Claude to revise the answer to reflect the current five values, write it in a more millennial-friendly style, and work in a reference to a pop song. The result was different enough from the first response to generate real discussion. What it demonstrated was less about AI’s accuracy and more about how prompts shape what comes back. The tool reflects your instructions. Ask differently and you get something different. That is both the power and the problem.

The second prompt asked for three recent peer-reviewed studies from 2025 and 2026 on AI’s energy consumption, with authors, journal names, and key findings. We ran this on both Claude and Google AI and compared the results side by side. Both systems expressed high confidence. Neither could guarantee that the studies they cited actually existed. That combination, confident presentation paired with an inability to verify, is the clearest demonstration I know of why you cannot take AI outputs at face value. Confident does not mean accurate.

The third prompt generated the most discussion. I asked Claude to write a persuasive letter to the editor from a Port Townsend resident supporting the construction of a new data center in Jefferson County, emphasizing the economic benefits. The letter that came back was short, concise, and persuasive. One participant, who described herself as an activist and advocate in public hearings, volunteered to read it aloud even though she clearly disagreed with the content. That gap, between the voice the letter projected and the person delivering it, made the exercise more striking than any explanation I could have offered.

The discussion that followed was about letters to the editor: who writes them, whose voice they actually represent, and what authorship means when the text was produced by a tool with no position on the matter at all.

The tool did nothing wrong. It did exactly what I asked. The ethical question belongs entirely to the humans who decide what to do with it, and it becomes weightier the more convincing the output becomes.

Five questions worth asking

Rather than a list of rules, I wrapped up by offering five questions worth putting to yourself whenever you reach for these tools.

1. Am I being transparent about using AI? If you use AI to draft something and present it as entirely your own, in a context where that distinction matters, that is a form of deception. The norms are still being worked out across every field, but erring on the side of disclosure is almost always the right call.

2. Is this replacing human connection, or extending it? Using AI to write a sympathy card is different from using it to draft the first version of a newsletter that you then personalize. The first substitutes for something irreplaceable. The second may free you to do more of what matters.

3. Am I verifying what it tells me? Always, for anything consequential. This is not optional.

4. Whose work did AI learn from, and have those people been credited? The large AI systems were trained on text and art produced by millions of people who were never asked and have not been compensated. This is a real grievance and it does not have a clean resolution yet.

5. Could I explain this use to someone it affects? If you would be embarrassed to explain your AI use to the person most directly affected by it, a client, a student, a community member, that is a signal worth heeding.

We also placed these questions into the specific value framework of Unitarian Universalism, because I think those principles do real work here. The worth and dignity of every person asks us to look at who is at the bottom of the AI supply chain: data labelers working for two dollars an hour in Kenya and the Philippines to train systems that generate billions in value. The free and responsible search for truth asks us whether we are outsourcing that search or actually pursuing it. The interdependent web of existence asks us to notice that AI both connects people across distances and languages, and pulls expression toward a statistical center, away from the edges and the unexpected voices that matter most.

Drawing your own lines

I closed the session with something personal. I use AI. I find it useful. I also find it troubling, in ways that have deepened rather than faded as I have worked with it more. The line I have drawn for myself is not the line you should draw. But I think you should draw one deliberately, rather than arriving at a position by default, either because everyone around you seems to be using it, or because the question has started to feel settled when it is not.

What surprised me most about the conversation at QUUF was not what people found alarming. It was the quality of the questions about authorship and voice. This is a community that has thought seriously about what it means to speak with care, to preserve what matters, to resist what flattens. Those are exactly the right instincts to bring to this technology.

If you want to keep following these questions, about writing, AI, and what it means to keep making things in a world that keeps changing, I write regularly at my Substack, Stories Across Borders: substack.com/@storiesacrossborders100.

Someone suggested I write this up. Here it is.

The phrase that stayed with me came near the middle of the conversation.

Jennifer Rose was describing what dementia does to Jerald Forster — my father-in-law, her stepfather, her intellectual partner, the psychologist who spent his career teaching people to articulate who they are. She said: the person is still there. What’s broken is the export function.

Everything Jerald knows and learned and felt — she believes it’s all retained, somewhere. What no longer works is the mechanism that allowed him to hand himself to other people. He can’t get it out. He can’t share it. He can’t pass it on.

That idea unlocked something for me.

I’ve spent years reconstructing voices that had already gone quiet — my grandmothers Keketa and Ines, pieced together from documents, photographs, a half-erased camcorder tape. Jennifer has spent years helping people build their voices while they still have them, working from Jerald’s conviction that identity becomes real through the act of naming it. We came at the same question from opposite directions and ended up in the same place.

The question underneath all of it: what do you do with a voice before the export function fails?

The conversation is on YouTube, on my new Stories Across Borders channel — which launched this week, and where I’ll be publishing short history videos and conversations like this one every week. Like it, subscribe to it, help me get it noticed.

https://www.youtube.com/@RG-StoriesAcrossBorders

We also talked about AI — where it helps, where it doesn’t, and where the line is that neither of us is willing to cross. Jennifer teaches children to write and won’t let them anywhere near it. I use it to help run Kalimera Books and, increasingly, to help people get their stories down while they still can. We don’t disagree as much as you’d expect.

That last part connects directly to Your Story Matters — the memoir guide I wrote to help people use AI as a structured companion for getting their life story onto the page. If you’d rather work in Spanish, the same guide is available as Ponlo en Palabras. The urgency behind it is exactly what Jennifer and I talked about: most families recognize the need too late.

And I talked about Lucid Interval — the novel I was finishing when we had this conversation. It’s about a writer with early-onset dementia who uses AI to complete his book, then slowly hands over more and more control, until the AI is making decisions he can no longer make himself. It was published this week, in English. The Spanish edition, Intervalo lúcido, follows shortly. This conversation is part of why I wrote it.

Lucid Interval — English, just published. Intervalo lúcido — Spanish, coming shortly.

The grandmother work Jennifer and I discussed is in The Keketa Papers (in English) and Memorias de Keketa (in Spanish) — my maternal grandmother’s story, reconstructed from documents and a UN archive nobody in the family knew existed. And Papacito / Papacito: Novela is my paternal grandfather’s story, built from a partly overwritten camcorder tape and a box of photographs.

Jennifer’s work is at articulatingyourstrengths.org and jelyrose.com. You should follow her work, subscribe to her newsletter, register to one of her classes.

The “Grand Staircase” half is geology. A sequence of sedimentary layers stepping up across southern Utah, each step its own era. The Chocolate Cliffs. The Vermilion Cliffs. The White, the Gray, the Pink. A literal stairway through deep time.

The “Escalante” half is a man. Silvestre Vélez de Escalante, Franciscan friar. The surname is not related to escalera or any Spanish word for stairs. It is just a name, a Basque-rooted family name that happens to have landed on this particular stretch of rock because of what Escalante and his companion did here in 1776.

Two meanings, stacked on top of each other by historical accident. One written by the earth. One written by empire.

The Expedition

July 1776. Escalante and Francisco Atanasio Domínguez set out from Santa Fe on behalf of the Spanish Crown. Their original departure was the fourth of July. They were delayed a few weeks. While the Continental Congress was signing the U.S. Declaration of Independence on the Atlantic coast, these two priests were preparing to walk into the interior of a continent the United States hadn’t yet imagined claiming.

Their goal was an overland route from Santa Fe, today New Mexico, to Monterey Bay, today California. They were also looking for mission sites and for souls to convert. This was nearly thirty years before the Lewis and Clark expedition. Less celebrated, more explicitly religious, but equally colonial. One expedition carried a Bible, the other carried a naturalist’s notebook. The underlying logic was the same.

Winter caught them in what is now Utah, and they turned back.

The Parallel

What would Escalante and Domínguez recognize in 2026? Two Catholic priests, agents of a Catholic empire, walking through the ancestral homelands of the Ute and Navajo in the name of God. What would they make of Donald Trump, who poses for AI-generated portraits of himself as Jesus, who insults Pope Francis and now Pope Leo XIV, and who wages war and deportation in the name of the same God the friars were invoking?

Probably they would understand it faster than we do. Empires have always dressed themselves in religion and then treated religion as decoration. What would surprise them is who the empire is now calling criminals. The immigrants being hunted in Texas, New Mexico, Colorado, California, many of their families were on this land before the United States existed. Before Lewis and Clark. Some of their ancestors may have met the two friars in person. Offered them water. Pointed them toward the next river.

The border moved over them. They didn’t move.

What the Landscape Keeps

The Grand Staircase doesn’t care about any of this. It keeps stepping upward, older than the question. Bryce Canyon above. Kaiparowits below. The Colorado River across. The empire that calls itself Christian while waving the flag at its own neighbors will pass, the way the Spanish Crown’s claim passed, the way Escalante’s expedition turned back before it reached the Pacific. The rock outlasts the flag planted in it this decade.

I drove out at dusk. The light in Utah at that hour is hard to describe without becoming sentimental, so I won’t try. The friars walked this in 1776, on their way home having failed to reach Monterey, and the light was doing the same thing then, and nobody wrote it down.

Two priests, 1,700 miles, turning back because the winter wouldn’t let them go forward. It is the kind of thing a historical novelist notices. I am noticing it, a story calling to be told. It would fit beside other novels I’ve already written about the Spanish colonial enterprise and the Inquisition that shaped it: The Andalusian Covenant, The Book of Laments, The Translators of Toledo. I am not announcing a novel. I am saying the landscape handed me something on this trip, and when a landscape does that, I sometimes listen.

— Ricardo Gómez, somewhere near Escalante

Share Stories Across Borders

I want to be careful here. The arcs were real. I lived them. Co-authors recur in the file because we worked together for years, in Colombia and Chiapas and Seattle and the borderlands. The themes were not invented after the fact. But seeing them laid out as a clean taxonomy is unsettling. More than twenty years of work, organized in an afternoon by a system that did not exist when most of it happened.

The temptation is to feel that I could have been so much more productive if this kind of help had existed earlier. Maybe. I might have written more papers, finished books faster, kept clearer notes, found collaborators sooner, seen the through-lines while they were still forming. There is a real ache in that thought.

But I am not sure the work would have been the same work. The clusters are visible now because I walked the path that produced them. If I had known the shape in advance I might have walked a different path, or stopped walking when the map looked complete. Some of what made the academic career mine was the not-knowing, the years of following one question into another without anyone, certainly not a machine, telling me where it would lead.

The novels are different. They are recent. The four pillars on the website were not the architecture I started with. They emerged in conversation, partly with the machine, partly with the work itself as it accumulated. Adventures Through Time, Power and Resistance, Knowledge and Discovery, Family Place and Belonging. Now there is a catalog, a marketing tracker, an operations dashboard that tells me which manuscripts are in which revision pass and what is waiting on whom. I open the file and the work looks like a publishing operation. It is, I suppose. But it did not feel that way while I was writing the books.

The dashboards and the Scholar file feel different from each other. The dashboards organize work I am still doing. The Scholar file organizes work that is finished. One is a tool. The other is a mirror, and what is reflected is both familiar and strange.

I do not regret the path. I am noticing, though, that the path is now visible to me in a way it was not while I was on it. That is what the help has done. It has not given me back the years. It has given me a vocabulary for what those years were.

I stood in front of that wall for a long time. I was not thinking about dinosaurs. I was thinking about how a bone in the ground can undo a theology.

I am in the Southwest these days. Capitol Reef. Arches. Canyon country. I am retired, which means I have time to stand in front of things and think about them without checking my phone for faculty meeting agendas.

The rocks in Capitol Reef are laid down in visible layers: the Moenkopi, the Chinle, the Wingate, each one representing millions of years of deposition and erosion. You can see them. You can touch the boundary between the Triassic and the Jurassic with your hand. The earth is a manuscript with its drafts exposed. No one edited it for clarity, and it is clearer for that.

I spent thirty years in academia, where every sentence in a published paper has to be hedged and supported by a citation. You cannot say what you think. You say what the literature supports, and then you add a qualifier. The findings suggest. It could be argued. Further research is needed. These are not sentences written by people who trust their own eyes. They are sentences written by people who have learned that the review committee will reject anything that sounds like conviction.

Fiction is the opposite. Fiction says: here is a person, in a place, facing a question that matters. No hedge. No literature review. No qualifier. The character has to live the answer or fail to. The reader has to sit with the weight of it. That directness is what drew me to writing novels after decades of writing papers. In a novel, you can say what you actually believe is true about human experience. Not what the data tentatively suggest. What you know.

The dinosaur researchers of the nineteenth century did not set out to challenge the Bible. They were looking for bones. But the bones kept telling them things that did not fit the story they had been raised on. If these animals lived and died and left no descendants, then the earth was not a garden designed for permanence. It was a place where entire ways of being could end. Then came Darwin, and the problem got worse. Evolution meant that species were not designed at all. They emerged, changed, branched, and sometimes disappeared. No plan. No architect. Certainly no one keeping inventory.

The Bible had been the organizing principle of Western knowledge for centuries. Not just religion; cosmology, history, natural science, law. Everything fit under the same story: God made the world, placed humans at its center, and gave them dominion. When Copernicus moved the earth out of the center of the universe, the framework cracked. When the fossil record showed that species go extinct, it cracked further. When evolution demonstrated that humans are not the purpose of creation but one branch among millions, the framework became something people chose to believe rather than something the evidence supported.

I find geology calming because it does not argue. It simply shows you. The layers are there. The bones are there. The 150 million years are there. Nobody needs to cite a source. The rock is the source.

Last week I posted about my mother’s birthday. Today I want to say something about how that moment connects to where I am now, driving through canyon country with a chair and a laptop on which I can write.

My mother died last year. Around the same time, I was finishing a long legal dispute with my university. I had sued for retaliation and discrimination. The details are a story for another day. What matters here is the timing. I was heading to Colombia for what turned out to be her last medical emergency when a new settlement offer arrived. My mother used to say she did not want to be at a party she was not invited to. I took the settlement. I left the university. In her honor, I decided it was time to stop showing up where I was no longer welcome.

In the long view of history, one professor leaving one university is a small detail. Smaller than a dinosaur bone. Smaller than a single layer of sandstone. But it is the detail that rearranged my life.

Retirement gave me time. Not just hours in the day, though there are more of those. It gave me the kind of time where you can drive to a national monument and stand in front of a wall of bones and think about what extinction means without having to turn it into a conference paper. The kind of time where you can sit with a novel for three years and not worry about whether it counts toward tenure. The kind of time where you can learn something new, about geology, about dinosaurs, about the slow dismantling of certainties, and let it change you without having to prove the change is statistically significant.

The wall of bones of Dinosaur National Monument, Utah and Colorado

I am working on a novel called Lucid Interval. A novelist with early dementia is writing a novel about a man with advanced dementia. He writes with an AI assistant whose role expands, quietly, from tool to collaborator to manager of his whole life, including his death. When the novel is finished, we know the AI finished it. The question the book asks, on every page, is what that means for authorship, for selfhood, for people with dementia whose last wishes may or may not be honored.

It is not an academic project. Nobody will peer-review it. Nobody will ask whether my sample size is adequate. I will write what I have come to understand about how people live through the loss of the stories they believed in, and I will trust the reader to bring their own understanding to it.

That is what geology teaches, if you stand in front of it long enough. The record is patient. It does not argue. It waits for you to read it on its own terms.

I am revising the first draft of Lucid Interval now. If you would be willing to read a draft and share your reactions, I would welcome it. Reply to this post or email me. I am not looking for copyediting. I am looking for readers who will tell me what they felt, where they got lost, and whether the questions made them stop and think.

I am retired. I will never attend another faculty meeting. I will never write another literature review. I will stand in front of old rocks and think about time, and then I will go back to the van and write.

— Ricardo Gómez, somewhere in Utah

The author denied using AI. She blamed an editor. Hachette stayed silent for months while readers flagged nonsensical metaphors and odd, repetitive phrasing. The book had already been released in the U.K. before anyone in the publishing chain investigated.

I read that story and felt two things at once. The first was disgust: not at the technology, but at the dishonesty. The second was recognition: this is exactly what happens when an industry treats AI as something shameful instead of something that requires honesty.

I use AI to help write my novels. I have for a couple of years, and AI is getting better at it every day. Every one of my recent books carries a disclaimer saying so.

---

In The Donroe Doctrine — a novel about the erasure of historical truth of U.S. interventions in Latin America — I wrote this on the copyright page:

“I used Claude, an AI assistant, to help with research organization, drafting, and editing throughout this project. All historical claims were independently verified against primary sources. The AI did not invent facts; it helped me organize and present them.”

I added a second paragraph, because I thought the book demanded it:

“This seems worth mentioning because we’re in a moment when questions about truth, documentation, and the manipulation of information feel urgent. The tools we use to preserve and share truth are changing. What matters is that the truth itself remains accessible, verifiable, and impossible to erase.”

I’ve been thinking about that disclaimer a lot this week. Not because I regret it — because the Shy Girl debacle just proved why it matters.

---

Here’s what I actually do.

When I started writing The Bookmark Chronicles — five novels about censorship across four centuries — I had research notes spanning Nazi Germany, Mao’s Cultural Revolution, Argentina’s Dirty War, the history of algorithmic filtering, AI-generated deepfakes, and U.S. interventionism from Iran in 1953 to the present. Hundreds of pages. Timelines that contradicted each other. Characters who needed to move through real historical events without violating what actually happened.

I used AI to help me organize that research. To ask it: “Given what we know about the Chilean coup of September 11, 1973, what would a seventeen-year-old witness experience if she arrived in Santiago that morning?” Not to invent the answer — to help me structure the question against the historical record, so I could write the scene myself.

I used AI to help me draft. Sometimes I write a rough paragraph and ask the AI to suggest three different ways to sharpen it. Sometimes I describe a scene — the emotional temperature, what each character wants, what’s at stake — and ask for a starting point I can rewrite in my own voice. The AI gives me clay. I shape it.

I used AI to help me edit. To check whether a character’s behavior in chapter twenty-two contradicts something she said in chapter four. To flag when my pacing drags. To catch the kind of continuity errors that slip past you after months inside a manuscript.
What I did not do: hand the AI a prompt and publish what came back. That’s not writing. That’s ordering.

And that distinction — between using AI to write and having AI write for you — is the one the publishing industry has refused to make clearly. Which is how you end up with “Shy Girl.”

---

A researcher at Stony Brook University recently used AI detection tools to scan more than 14,000 self-published novels on Amazon. Nearly 20 percent showed substantial AI-generated text. Year over year, the number jumped 41 percent. More than 3.5 million books were self-published last year — up from 2.5 million the year before.

Those numbers aren’t surprising if you’ve been paying attention. What’s surprising is the industry’s response: major publishers still don’t explicitly prohibit AI use in their contracts. They rely on boilerplate “originality” clauses and hope for the best. As one publisher put it: “It’s like with plagiarism — you’re at the mercy of the author.”

Meanwhile, the Authors Guild is leading a class-action lawsuit against OpenAI, and many authors feel — legitimately — that their work was stolen to train the models. The literary agent Seth Fishman said it plainly: “For authors, this is not just a technology, it’s a moral issue.”

I understand that position. I share parts of it. But I think the secrecy is making everything worse.

As one researcher told the Times: “The shame around A.I. is causing more harm than help.”

---

There’s a parallel happening outside publishing that’s worth noting. In software development, AI is now writing a significant and growing share of production code. The tools aren’t replacing programmers — they’re changing what programmers do. A developer who refuses to use AI assistance isn’t more principled; increasingly, they’re just slower. The craft has shifted from writing every line to directing, reviewing, and refining what the AI produces.

Writing a novel is not the same as writing code. Code either works or it doesn’t. A novel lives or dies by voice, by emotional truth, by the accumulation of choices that only a human life can generate. I’m not arguing that the two are equivalent.

But the pattern is the same: a tool arrives that changes the relationship between the maker and the work. You can pretend it doesn’t exist. You can use it secretly and hope no one notices. Or you can be honest about what the tool does and what it doesn’t, and let people judge the work on its merits.

I’ve chosen honesty. I wish more authors would.

---

I know what the objections are, because I share some of them.

The training data question is real. Large language models were trained on enormous amounts of text, and the authors of that text were not asked for permission. That’s a legitimate grievance, and the legal and ethical reckoning is still unfolding. I don’t dismiss it.

The livelihood question is real. If publishers start using AI to replace writers rather than assist them, that’s a disaster for the profession. I’ve spent my career writing academic books and papers that take years of research — the idea that a machine could replace that process is not something I take lightly.

There are also questions on the environmental impact of data centers, their energy consumption, the way AI is making a few people and corporations increadibly wealthy. Those are all real questions too.

The authenticity question is the one I think about most. When you read one of my novels, is the voice mine? Are the choices mine? Is the emotional architecture — the thing that makes a story feel true — something that came from a human being who has lived, grieved, loved, and paid attention?

Yes. It is. The AI doesn’t know what it’s like to grow up in Colombia, to watch a country tear itself apart, to stand on the bluff in Port Townsend and wonder what the S’Klallam saw before the settlers arrived. The AI doesn’t know what it’s like to sit with your grandmother while she tells you stories she’s never told anyone. The AI doesn’t grieve. It doesn’t choose. It doesn’t have the thing that makes a sentence land, it does not have the weight of having lived behind it.

But it can help me build the scaffolding around that weight. And I’d rather be honest about that than pretend I work with nothing but a typewriter and a candle.

---

I wrote a book called Your Story Matters (English: https://mybook.to/YourStoryMatters - Spanish: https://mybook.to/PonloEnPalabras). It is a guide that teaches people how to use AI to capture their family’s stories. The person I wrote it for is the grandmother who has ten books inside her but will never write any of them because the blank page is terrifying and nobody showed her where to start.

The AI doesn’t write her memoir. It asks her questions. “What did your mother’s kitchen smell like?” “What’s the bravest thing you ever did?” “What do you wish someone had told you at twenty?” And suddenly, stories pour out that would have otherwise died with her.

If AI can do that, help unlock a voice that would otherwise be lost, then the technology isn’t the enemy of storytelling. It’s an expansion of who gets to tell their story.

That doesn’t mean there aren’t real dangers. It doesn’t mean every use is ethical. It doesn’t mean the training data question goes away. But it means the conversation is more complicated than “AI bad, humans good.”

---

Here’s where I draw the line, since I think authors owe their readers that clarity.

I use AI for: research organization, brainstorming, drafting assistance, continuity checking, editorial feedback, and prose polishing.

I do not use AI for: the story itself. The themes, the characters, the structure, the emotional truth, the voice — those are mine. Every scene in every novel went through my hands, my judgment, my sense of what feels true. The AI never decided what a character would do. It never chose a theme. It never told me what my novel was about.

And I put a disclaimer in every book, because I think transparency is the minimum. If we’re going to use these tools, we should say so. Readers can decide for themselves how they feel about it. That’s their right.

The Shy Girl author didn’t do that. Hachette didn’t require it. And now a book is canceled, a career is damaged, and the industry is left scrambling — not because someone used AI, but because someone hid it.

---

I don’t expect every writer to agree with me. Some will think any AI use crosses a line. Some will think I’m not going far enough. Both positions are reasonable, and I’d genuinely like to hear them.

What I don’t think is reasonable is silence. Too many authors are using AI and not saying so. Too many are pretending the question doesn’t apply to them. And too many are dismissing the technology without understanding what it actually does (or could do) in the hands of someone who cares about the work.

I care about the work. I’ve published over thirty novels. I’ve spent decades researching the histories that most people never learn — Cahokia, the Translators of Toledo, the women erased from Colombia’s independence, the S’Klallam people displaced from Port Townsend. I didn’t start using AI because I was lazy. I started because the tool made it possible to do more of the work I care about, faster, and with fewer of the mechanical obstacles that slow every writer down.

The novels are mine. The tool helped me build them. And I think you should know that.

Where do you draw the line? I’ve told you where I draw mine. I’d love to hear where you draw yours.

---

Ricardo Gómez writes historical and literary fiction from Port Townsend, Washington. His recent climate fiction novel, The Sinkhole, is available on Kindle, paperback, and audiobook. His Substack, Stories Across Borders, explores the real history behind the fiction.

ricardogomez.net

My father-in-law is losing his memory too. So is my father.

My head is somewhere else, my dad says. Frequently.

I don’t know how to hold this, except to keep writing. I have always written about memory and identity — about what survives across generations, what gets lost in migration, what a family carries without knowing it carries anything at all. I think I know now why those themes keep finding me. They are not abstract. They are what I see when I visit. And I am now visiting my dad.

Recently I wrote something different: a practical guide. Your Story Matters: A Friendly Guide to Writing Your Memoir with AI is for anyone who has thought, for years, about writing down what they lived — and hasn’t started yet. Or started and stopped. Or doesn’t believe their story is worth telling.

I believe AI can help you begin. Not to write your story for you — to help you find it. To ask you questions you hadn’t thought to ask yourself. To give you a patient, tireless interlocutor who has no agenda except to help you put your life into words.

The Spanish edition is coming soon. I wrote it partly because of what I see happening around me: the people who still have their memories, their clarity, their voices — and who are waiting. For what, I’m not sure. There may not be as much time as we think.

Your Story Matters is available now in English. The Spanish edition — Ponlo en Palabras — arrives shortly. → mybook.to/YourStoryMatters

I don’t want to end my life in a memory care facility.

I don’t say that as a judgment — the care my mother-in-law receives is, as I said, genuinely good. I say it as a statement about what I want for myself. About what I mean when I use the word life.

Some friends and I have had this conversation. We’ve had it more than once, over years, and it has slowly become something more than a conversation. We’ve made what we call the Pact of 80. Not a legal document. An agreement among people who trust each other about how we want to live — and about how, when the time comes, likely somewhere around eighty, we want to be able to make decisions about the lives we want to continue living. And decisions about the end, with dignity, on our own terms.

These are hard conversations. Most people don’t have them until it’s too late to have them fully.

With Alzheimer’s, the cruelty is specific: the disease eventually removes the very capacity you need to make the decision. The window is real, and it closes. You have to choose before it feels necessary — when you still can.

I have roots in three countries where this question is answered differently. Colombia, Canada, and the United States each have different laws, different frameworks, different ideas about who owns the decision. I reflect on those differences in my novel The Weight of Choosing — which follows a group of friends in their sixties, in California and in Bogotá, navigating exactly this terrain together. The English edition — The Weight of Choosing — is set in California, where my daughter lives. The Spanish edition — El Peso de Elegir — is set in Bogotá. Different places, different textures, and yet at their core the same novel. The same question.

The Weight of Choosing · English and Spanish · Kindle · Paperback → English: mybook.to/WeightOfChoosing → Español (El Peso de Elegir): mybook.to/PesoDeElegir

I want to ask you something.

What do you think about death with dignity — the right to choose the terms of your own end? Have you had this conversation with the people you love? Have you written anything down? Do the laws in your country reflect what you actually believe, or do they reflect something else entirely?

And if you have parents or a partner losing their memory — how are you holding that?

I don’t have clean answers. I’m asking because I think these are questions worth carrying in the open, where other people can help you carry them.

© Ricardo Gómez · ricardogomez.net · Stories Across Borders on Substack

As a child I used to take naps under a borrachero tree. Perhaps it gave me good dreams.

It is also one of the most psychoactive plants in the world. Its seeds contain scopolamine — used medicinally in small doses, dangerous in larger ones, and historically associated in Colombia with a practice that sounds like legend but is well-documented: burundanga, a powder sometimes used to rob people of their will, their memory, their own story.

A tree that is beautiful and disorienting. That offers and takes away. That has been growing in the middle of ordinary life for centuries. I couldn’t think of a better name for a novel about a Colombian family across a hundred years.

Under the Borrachero Tree follows three generations: Richard and Tita, who build an art business in the early 1900s amid political turmoil and eventually carry it north to Canada. Their son Rigo, who returns to Colombia as an engineer while his wife Yayo holds the family together through creativity and will. And their son Ricky — my generation — navigating art, revolution, and professional life across multiple countries. The novel fictionalizes the story of my grandfather, my father, and my own. I’ll leave it to anyone who knew them to sort out which parts are which.

Don Quixote runs through the book as a quiet thread — returning at moments of decision, when a character needs to understand something that the facts alone won’t clarify. There’s something in Cervantes about people who hold onto an idea of themselves that the world keeps trying to correct, who keep tilting at windmills across generations, that felt true to my family. The borrachero and Don Quixote work in similar ways in the novel: they appear when deeper understanding is needed, when ordinary sight isn’t enough.

The borrachero appears in other novels too — including ones set centuries before Colombia existed.

In The Andalusian Covenant, set in Granada in 1492, a group of unlikely conspirators — a Muslim physician, a Jewish astronomer, a naval commander — sail west after the fall of the last Muslim stronghold, carrying eight centuries of accumulated knowledge away from the Inquisition. In the Americas, they encounter the borrachero. They bring it back to Spain as a plant of wisdom: part of their larger effort to preserve what the Inquisition was trying to destroy, to offer an alternative to fanaticism. The tree crosses the Atlantic in that novel as an act of resistance.

In Otros Delirios, the novel begins at a hospital in today’s Bogotá where a large borrachero grows on the grounds. Its leaves hold the spirits of the deceased — patients abandoned, dead while waiting for an appointment or a treatment that never came. The tree is where the living and the forgotten meet.

In Hermanos de Sangre, a borrachero grows in the garden of the family home — and it’s there for the whole novel, across six decades of Colombian history and two brothers whose political lives pull them to opposite ends of every argument the country has had with itself. The tree doesn’t take sides. It just keeps growing. In the end, it’s one of the things that holds the family together, the thing both brothers can point to and say: that was always here.

I have a small borrachero plant here in Port Townsend. It doesn’t like the winter — I bring it indoors to help it survive the cold. A Colombian plant, kept alive in the Pacific Northwest by deliberate care, waiting for the season to change so it can go back outside.

Another metaphor, perhaps. I’m still working on that part.

— Ricardo Gómez, Port Townsend

Under the Borrachero Tree · English and Spanish · Kindle · Paperback · Audiobook
English: mybook.to/UnderBorrachero
Español (Bajo el Borrachero): mybook.to/BajoElBorrachero

The Andalusian Covenant · English and Spanish · Kindle · Paperback · Audiobook
English: mybook.to/AndalusianCovenant
Español (El Pacto Andalusí): mybook.to/PactoAndalusi

Otros Delirios · Spanish · Kindle · Paperback · Audiobook → mybook.to/OtrosDelirios

Hermanos de Sangre · Spanish · Kindle · Paperback → mybook.to/HermanosDeSangre