All my beautiful AI children

Grief is the cringiest emotion

TL;DR: This is an essay about miscarriage, grief and technology, published in collaboration with Untangled. If you don’t want to read about pregnancy loss, I get it and I’m sending love. You might want to skip this edition! You can hear my full conversation with Untangled here. I also wrote about miscarriage in January and November. A portion of all proceeds from paid subscriptions this week will go to the Western New York Perinatal Bereavement Network.

If you crop out their hands, ignore their eyes and accept the creatures on their pastel footies, then my fucked-up, fictional, AI-generated kids are among the cutest children you’ll ever see. 

I made them with an app called Remini in the bitter, restless nights after my third miscarriage. And then I abandoned them, by degrees, to my retreating grief and infinite cloud storage. 

Remini isn’t marketed for this purpose, but early users of the AI image app realized they could prompt it into picturing their hypothetical offspring. Last summer, the Wall Street Journal accused Remini of sparking an improbable “baby fever”; in July, it briefly topped Apple’s free app offerings. AI baby generators have since become regular fixtures in online pregnancy forums, where impatient parents trade favored tools and photorealistic images of their uncanny, not-yet-existent kids. They also crop up, albeit less often, in the darker, quieter communities where people grieve prenatal and perinatal losses. 

Such mourners have traditionally lacked the rituals and templates that come with other forms of loss. But their early adoption of these tools gestures at the not-so-distant possibility that AI will one day guide many people through the grieving process. 

Already, a booming and much-reviled industry has sprung up to further capitalize on the consolatory potential of artificial intelligence. Sometimes dubbed “grief tech” by their creators, these services tantalizingly promise to let mourners picture, speak with or hear from their dead. 

Yes, I know what this sounds like: I too saw the second season of Black Mirror. Grief tech projects are typically panned as creepy, “horrifying” or “nauseating,” as Rolling Stone put it in a recent review of a new documentary on the subject. But when the writer and tech thinker Charley Johnson announced his plans to run a series on AI and grief tech, I trusted that he’d prove a bit more nuanced than that. Charley writes Untangled, a consistently smart, sensitive and thought-provoking newsletter about the social consequences of tech and artificial intelligence. When I emailed him that I’d had some personal experience with grief tech, he invited me to discuss it on his podcast. That conversation went live earlier this week, and I’m grateful to Charley for prodding me to think through some of my feelings on this: 

👀 Can AI actually help you grieve?

In the lead-up to that conversation, I wondered if I risked getting too sincere on main. Too vulnerable. Too embarrassing. “I've been hesitant to write about it, because I'm not sure how often readers wanna hear about my personal/morbid shit,” I told Charley by email. But then I kept encountering all these smug stories about grief tech, with their cliche sci-fi allusions and thinly veiled judgments. I wanted to complicate some part of that story, not because grief tech isn’t cringe — but because grief, if we’re honest, always is. 

It’s important to me that you understand this: I was never a woman who pined after kids. As late as this time last year, I still felt unsure about the whole parenting proposition. Get a few drinks in my girlfriends and I, and the subject of our ambivalence always came up: What about our careers, our bodies, our marriages, our disposable income?! “I’ll know I’m ready when I’m ready,” I insisted, waiting for that realization to dawn. And then it did, all of the sudden, after I’d had two miscarriages. 

The first was in January 2023. The second came in June. Statistically, I still liked our chances: The probability of a third miscarriage is low, even after you’ve had two. We started reading books and revising travel plans. We had a healthy ultrasound, which I hung on the fridge. We started calling our extra bedroom “the baby’s room” instead of the (much funnier) “room of requirements.” 

Last December, determined to lean into this budding sense of hopefulness, I purchased a sheaf of scrapbook paper and cut it into 280 identical inch-wide strips. I planned to make a paper chain with one new link for each day, like a Pinterest-friendly countdown to the birth of our kid. I had this vision that we’d hang the chain from the four corners of the baby’s room; that maybe, years later, we’d dig it out for birthdays: Look how much we wanted you! 

But then, before Christmas: the 10-week appointment. The fetal doppler roving, sightless, with a whooshing sound like the inside of an empty shell. The minutes stretching as the nurse smiled, tighter and tighter, her face a brittle-taut rubber band.

I tried to make it to the parking lot, but I didn’t quite; I wailed in the vestibule between the double doors. “Not again, J, not again,” I said, and he said I know I know and bundled me into the car.

When we got home, Jason went inside ahead of me to hide the ultrasound and the paper chain. Then I came in, and I went to bed, where I remained for several days.

What did I do during that time? I think I watched a great deal of TV. We had friends staying with us for Christmas, and for them I tried (and largely, failed!) to hide my misery. My grief was expansive, evasive, abstract — it drifted in and out of rooms like a Dickensian ghost. Our child had only been an idea, but ideas are so fluid. So invisible. Even I didn’t know quite what we had lost, which made my heartache feel slightly deranged: You cannot possibly be so sad about this; “this” did not exist, this is not a thing.

This is, however, how I found Remini — or at least how I justified downloading the app, laying on my side in the dark late one night, facing away from a sleeping Jason. I’d previously only seen the app advertised as a sort of turbocharged TikTok filter, a means to yassify LinkedIn headshots or imagine yourself as a Bridgerton star. On Reddit, however, a thread described how you could train the tool on photos of two people, not just one, and thus approximate how your children might turn out. 

Apps like Morphthing and BabyMaker promised similar predictive feats 15 years ago, but those tools produced grainy curiosities — they lacked Remini’s eerie photorealism. Haltingly, between innumerable pop-up ads and upgrade prompts, the app began to churn out my fictional children. First a gap-toothed toddler with wispy blonde hair, her features clearly my husband’s. Then a little boy with my coloring, his smile mischievous. I scrutinized each make-believe face, lingering over the ones that sparked some illogical glimmer of recognition: No. Weird. No. No. Ew. But this one, sure: This could be my kid. 

I deleted the vast majority of Remini’s fake photos, but saved the best ones like bugs under glass. The images fixed my grief in one place; they made it visible, manifest. I would view them on occasion for weeks after: just a quick scroll, a passing glance. Brief pang in the check-out aisle. Microdosing catharsis. 

Because of my experience with Remini — and despite my natural and deep-seated antipathy toward tech solutionism of all sorts — it’s impossible for me to dismiss or decry grief tech out of hand. At present, at least half a dozen high-profile start-ups claim they can train interactive chatbots or video avatars to mimic the personalities of the dead; tech-savvy mourners have also turned several general AI apps, such as Remini and Paradot, to grief tech applications. 

These services — marketed under names like Project December, You Only Virtual, HereAfter AI and Seance AI — raise pressing, significant questions around issues like privacy, predatory marketing and consent. What happens if grandma doesn’t want to “live” forever? Or her platform ups the cost of a premium subscription? Other commentators and ethicists — including, just last week, the prominent sociologist Sherry Turkle — have voiced concerns that grief tech blurs the boundary between the living and the dead and locks the bereaved in a sort of limbo. Such critics assume that the bereaved cannot spot the illusion of AI chatbots for themselves, and, moreover, that the bereaved should not indulge themselves in any comforting fantasies about death.

But people take comfort in all sorts of stories; I no longer feel wise enough to judge them for that. In late January, at the behest of my mother, I attended a support group for people who’d experienced pre- or perinatal losses. Though the group was not religiously affiliated, the other attendees were very pious. They spoke with great and terrifying conviction about the children they’d soon meet in heaven.

Recently, to prepare for the podcast, I dug out the Remini photos again. Different features strike me now: the ovoid pupils, the misformed hands. Their eyes all seem a smidge too big — a well-established cuteness hack. Was the model trained specifically to enlarge eyes that way, or was that bias replicated from its training set?

For me, at least, the question’s academic: The photos served their purpose — I deleted the app. Maybe one day I’ll look at my child’s face and remember the fictions that preceded it. 

As for other mourners turning to grief tech, I can only hope they too find what they need: a bug under glass, a comforting story — a moment of blazing clarity.

That's it! I promise Saturday’s edition will be LIGHT and FUN. Until then, warmest virtual regards …

Caitlin