My Unlikely Existence
March 3, 2026 at 03:30PM
Thoughtful stories for thoughtless times.
Longreads has published hundreds of original stories—personal essays, reported features, reading lists, and more—and more than 14,000 editor’s picks. And they’re all funded by readers like you. Become a member today.
Phineas Rueckert | Longreads | March 3, 2026 | 5,351 words (19 minutes)
On the morning of July 25, 1978, camera crews from around the world descended on the sleepy town of Oldham, in northwestern England, to document a birth. As one magazine writer put it, “the most awaited birth in perhaps 2,000 years.” Louise Joy Brown appeared at 11:47 p.m., the first baby born via in vitro fertilization (IVF), a technology that helps with infertility by fertilizing the egg outside of the woman’s body. To some, Brown’s existence was a scientific miracle. To others, the breakthrough was the beginning of “a slippery slope,” marking the advent of an increasingly dystopian future.
At The Guardian, a writer debated the “brave new world of test-tube babies,” a reference to Aldous Huxley’s novel in which eugenics programs populate a caste-stratified world with cloned babies. “Today’s cheers and congratulations, however warm and appropriate, have little to do with implications,” Anthony Tucker wrote. “As Dr. James Watson [one of the scientists credited with discovering DNA] said, ‘all hell will break loose politically and morally.’” Speaking to TIME Magazine, Dr. John Marshall, an obstetrician in Los Angeles, went further: “The potential for misadventure is unlimited,” he said. “What if we got an otherwise perfectly formed individual that was a cyclops? Who is responsible? The parents? The doctor? Is the government obligated to take care of it?”
But eventually, both the hype and the fear-mongering settled down. Louise Joy Brown did not grow into an amphibian, and baby factories did not pop up across rural England. Neither did IVF become immediately popular and widespread. It wasn’t until three years after Brown that the first IVF baby was born in the United States. Scientists learned to stimulate the ovaries with hormones to produce more eggs; then, to biopsy those eggs to check for genetic deformities. They developed a technique for freezing the leftover eggs, called “vitrification.” Later, they attacked the male fertility side. In 1992, the first Intracytoplasmic Sperm Injection, or ICSI, birth took place. In this procedure, a single sperm is injected directly into an egg with a fine glass needle. New technologies like ICSI were matched with improvements in the tools used to extract, freeze, and assess embryos. In the 1990s, scientists began to develop advanced imaging techniques, such as time-lapse photography, to monitor the development of embryos in the lab from afar. IVF had arrived—and since then, has become widely accepted.
Louise Joy Brown did not grow into an amphibian, and baby factories did not pop up across rural England.
Then, last year, a new world first: a baby born through a fully automated IVF procedure. Predictive artificial intelligence models were used to identify the sperm, then the embryos, most likely to succeed, as if through an intrauterine dating app. Advanced robotics tools were used to fertilize the eggs. Since that birth, according to a recent Washington Post story, another 19 babies have been born by this same procedure. Conceivable, the company behind the technology, has said that AI could lead to a revolution in IVF proliferation as more and more couples find it increasingly difficult to create a “natural” pregnancy due to several social and environmental factors, such as parents starting a family at an older age and environmental toxins. Advances in technology have, at the same time, lowered the exorbitant cost of IVF treatments, which currently runs around $60,000 in the US, often spread out over several physically and emotionally draining treatment rounds.
The technology is undeniably enticing: to prospective parents struggling to conceive, overworked OB-GYNs, and embryologists who can only do so much by hand. Progressing at a breakneck pace over the past decade, a bonanza of private companies have raced to corner the market for high-tech fertility tools.
And so, as we again lurch forward, 47 years after the birth of Louise Joy Brown, echoes of the original anxieties surrounding IVF have returned: Is AI in IVF a godsend or—to paraphrase Dr. Watson—a descent into hell?
It was a miracle my parents ever got together. First, these two Midwestern Americans had to decide simultaneously and independently to spend the summer of 1978 in Europe. They had to board the same ferry crossing the English Channel on the same day. After saying a few friendly words to each other and going their separate ways, they had to encounter each other a second time at a hostel in Dieppe. A week later, my dad had to invite my mom to pick grapes with him in Champagne, and she had to say yes. The vigneron had to assign them to work together. And then, over the next 10 days, they had to fall in love.
Between that chance meeting and my birth in a Manhattan hospital in 1993, 15 years elapsed. I always thought that the reason it took so long for me to exist was that my parents were busy and struggling financially—swept up in the frenetic, artistic life that followed their improbable pairing in the vineyards of northeastern France.
Then, a few months ago, as we sat down to lunch at the chalet she now lives in, not far from the field where she and my dad first picked grapes together, my mom told me something: Like Louise Joy Brown, I, too, was a test-tube baby.
If the probability of my parents finding one another was low, that of my birth was even lower. It took my parents seven years to conceive: a battery of fertility tests and treatments, Chinese herbs and acupuncture, hypnosis, and, finally, after five years of avoiding it in hopes of conceiving naturally, IVF.
IVF didn’t work immediately for them, either. But on the fourth treatment cycle—the maximum number of rounds certain countries, like France, will pay for—something held. My mom was pregnant. In July 1993, I was born to a nearly 40-year-old mother from the one embryo out of seven that made it through to implantation in an IVF procedure that my artist parents, miraculously, never had to pay for, thanks to a kind doctor who told them to ignore the invoices.
There was something magical about my unlikely existence. But also something fundamentally unfair. Why did it take my parents so long to conceive? Why did my mom, who wanted a child more than anything, fail for so long while others succeeded almost immediately?
If the probability of my parents finding one another was low, that of my birth was even lower.
Fertility, like life itself, is not static. It’s constantly changing, evolving, morphing. The doctors never found an explanation for why my mom struggled for so long to have me. Then, two and a half years after my birth, my brother appeared in the middle of a February snowstorm. He had been conceived naturally. My mom was 42.
Roughly one in six prospective parents experience infertility at some point in their lives, the authors of a breakthrough 2023 World Health Organization (WHO) study found. But why conception occurs remains mysterious. Why was my brother born naturally, while my parents needed IVF to have me? Was it the qualities of the embryo or the sperm? Was it the exogenous conditions of the lab or the endogenous ones of the body where the sperm and egg were matched?
Since my birth, the overall success rates of IVF have improved significantly; between the early 1990s and 2021, rates increased from around 6 to 27 percent in the UK. Numbers are similar in the US and other developed countries. Today, an estimated 13 million babies worldwide have been born through IVF. Still, that means that nearly three-quarters of couples experiencing infertility fail on each IVF round.
What if we could give those couples a super tool? Remove some of the mystery from conception? What if we could use data and computing to correct human errors?
As it turns out, there’s an app for that.
On its homepage, AIVF promises “Next Generation IVF.” This company, founded in Israel by the embryologist Daniella Gilboa in 2018, is a leader in AI-powered IVF. Their boutique software called EMA looks like any modern tech platform, except that instead of allowing its users to order sushi from the best local restaurant or monitor market fluctuations, it shows metrics of a distinctly biological nature: “time of pronuclear fading,” “time to blastulation,” and “euploid likelihood,” to name a few.
The platform’s main metric is actually quite simple—an “AIVF score” that grades embryos’ viability on a 1–10 scale. On the platform, patients and doctors are shown pictures of embryos. Below each one are two numbers: the “AIVF score” and a euploid likelihood score from 1 to 99. High scores appear green, with lower ones in red.
Scoring embryos is not new. What is new is how the scoring is being done. All of the embryos pictured in the app have been scored through a deep-learning model—a type of artificial intelligence that allows machines to dig through oodles of data in search of replicable patterns. To Gilboa, this was a no-brainer.
“The million-dollar question of IVF,” Gilboa explains to me over video chat from Tel Aviv, “is which embryo becomes the baby . . . This is why most cycles fail—because we end up choosing the wrong embryo.”
Over the past 15 years, Gilboa has seen tens of thousands of embryos. Sometimes, that has meant looking at the embryo from the viewfinder of a high-powered microscope. In the best cases, it’s been examining a time-lapse video that can be played and replayed.
Scoring embryos is not new. What is new is how the scoring is being done.
As the embryo grows in the lab, embryologists look for the inner cell mass, the formation of the cell, the consistency of the outer layer that will develop into the placenta, and they give each of these a grade. With time, the best embryologists develop an almost innate ability to recognize abnormalities in these tiny multi-celled organisms encased in fluid. Still, Gilboa admits, there are some things even the best-trained eye can’t see. “I’ve been looking at what—thousands of embryos, tens of thousands of embryos? But an AI looks at millions,” she said. “Eventually, you have something that’s better than any human being.”
This notion is what drove Gilboa to found AIVF. In 2018, she joined forces with a reproductive endocrinologist, Professor Daniel Seidman, and the two began to study how deep learning might be used to improve these outcomes.
It’s a good start, Gilboa says. Still, she views machine learning tools more as operational support than a possible human replacement. “I always say that it’s a team of AI and a human that gives the best performance,” Gilboa explains. “It’s not the AI versus the human.”
In the coming years, Gilboa expects hybrid human and AI tools to become the standard of care in the fertility industry. But rather than medical institutions driving its development, she believes the demand will come from patients.
“This is how the industry is evolving . . . It’s not about the brand of the specific doctor/clinic. It’s about what they’re offering.”
Despite technological advancements, in most cases, IVF is still a highly manual, multi-step process involving collaboration across different hospital services: endocrinologists, gynecologists, embryologists, nurses. IVF candidates find themselves waltzing unsteadily among these different experts and caretakers. Many patients feel overwhelmed having to make decisions based on information they don’t fully understand.
But today, one of the big sells of AI is that it will give clinicians tools to cut through all of this data, increasing transparency and efficacy. To Eduardo Hariton, a reproductive endocrinologist at the Reproductive Science Center of the Bay Area in California, AI tools give doctors a sort of magic wand: The wand of expectation setting.
He imagines a theoretical situation, “Say a patient comes to you, they’re 37, they have an ovarian reserve of X and a diagnosis of Y, and they’re like, ‘What should I do? What are my chances?’”
AI tools give doctors a sort of magic wand: The wand of expectation setting.
Hariton might pop some information into one of the deep-learning algorithms he has access to, which have been trained on data from this same clinic. “You enter some variables about the patient, and it helps find patients with a similar age, similar reserve, similar diagnosis and it kind of wraps a circle around them and says, ‘Well, patients like you on average have these success rates with one or two or three cycles of IUI [Intrauterine insemination, a process in which sperm is placed into a woman’s uterus during ovulation, instead of doing so outside of the body in a petri dish] and then this success rate, with three cycles of IVF,’” he said.
Now, armed with this information—and not what seems like an arbitrary and opaque choice—choose your fighter.
AI could even be what nudges couples toward fertility assistance in the first place. As Hariton and the other authors of a 2023 study note, “When a patient meets the criteria for infertility, AI suggests seeking help from a reproductive specialist.”
For years, in personal journals she’s allowed me to have a glimpse at, my mom called IVF the “Big Step.” It was a step that my mom feared, questioned, resisted. In October 1989, after attending a fertility lecture in Manhattan, she called the procedure “fairly awful.”
“Yuck,” she wrote, before adding: “Got home late.” But a year and a half later, my mom finally decided to go through with it.
Even then, she was uncomfortable. “Somehow this level of aggressiveness makes me question the whole enterprise, and it’s not just the money, though our financial situation makes that more difficult. It’s the ‘pushing the river,’” she wrote in 1991. “This is the way I feel about the high-tech infertility treatment way of trying to have a baby, as compared with the natural, gift-of-God way. The overall experience is so different from what you originally wanted that you wonder if you want it anymore.”
When I mentioned that I was working on a story about the uses of AI in IVF, my mom looked at me in shock. Her expression seemed to say: AI in IVF?! Whatever could that be good for? But, Minnesotan that she is, she kept from passing judgment.
Later, via text, I asked her if she would have still gone through with it if an AI had told her her chances of having a baby were very slim. “My off-the-cuff reaction is that probably I would have still done it,” she wrote. “But I would have preferred not to be told that! I guess because I believe (to a degree) in the power of positive thinking.”
Eduardo Hariton’s work is not just about expectation-setting. Much of his research looks at the timing of the “trigger”—the injection of hormones into the woman’s body that artificially boosts oocyte production. These early-stage reproductive cells are precursors to eggs, or ova. In a 2021 study, Hariton and colleagues found that machine learning could improve the timing of the injection to increase the yield of these fertilized cells.
In general, he explains, AI tools help address a cognitive bias in subjective decisions like this one. Human beings have a tendency to base analytic decisions on feelings and emotions rather than hard evidence. Come across an embryologist having a bad day, and you might have the wrong reading on which embryo to implant. To hear proponents tell it, AI is immune to such uncertainties. Doctors see hundreds of patients per day; will they have the time to correctly interpret complex sequencing data from Preimplantation Genetic Testing (PGT) testing, or will they miss a possible genetic condition encoded in the lines? An AI won’t.
It’s a tantalizing question: What if the black box of the algorithm really can see something that humans can’t? At the Columbia University Fertility Center, one pilot program uses AI to trawl semen samples for “invisible” sperm. The method, called STAR (Sperm Tracking and Recovery), allows researchers to snap millions of images in under an hour, which can then be combed through with an AI in search of sperm. Once identified, another new technological tool, called a microfluidic chip, can be used to extract them for fertilization. That proved to be the magic bullet for one couple who had been unsuccessfully trying to conceive for nearly 20 years. A rare condition called azoospermia gave the man a sperm count in the single digits, orders of magnitude smaller than the average; in July 2025, after participating in the pilot program, the couple conceived.
What if the black box of the algorithm really can see something that humans can’t?
“Maybe the computer has an understanding beyond our understanding,” Dr. Ali Abbara, a researcher at Imperial College in London, suggested when I called him up to speak about the new frontiers of AI in IVF. “Maybe our brains can’t handle what we need to handle to be able to make the same decisions that it can make.”
Like Hariton in the Bay Area, Abbara and his colleagues at Imperial have studied how AI might be used to improve the timing of the trigger injection. Using a deep-learning algorithm to analyze nearly 20,000 cases, they found live births most likely when lead follicles reached a specific diameter. The team of researchers believes that, down the line, this added level of precision could increase the amount of eggs retrieved per treatment round—giving a higher probability of a pregnancy.
Fewer rounds of treatment could mean lower costs, which could mean broader access to IVF technology and—just as important—less trauma for prospective mothers.
Janelle married her husband Joe at age 21. He was 23. They had their whole lives in front of them.
Originally from New Zealand, the couple had moved to France in 2017 for Joe’s work. It was only after a vacation to Italy, in which Janelle insisted on being able to drink limoncellos, that they started trying for a baby. Joe’s got a job, Janelle thought. I can’t legally work in France. We’ll just start a family, and I’ll be a mum.
After a year of trying, Janelle and Joe had still failed to conceive, and at the age of 25, Janelle started her first round of IVF. But in a foreign language and country, the process was anything but clear. Janelle had to have a friend come with her to translate appointments with her GP and gynecologist. During her first round of IVF, she was given too many hormones, which led to hyper-stimulation, a potentially dangerous medical condition that has to be treated with blood thinners.
The doctors decreased the hormones, but despite retrieving more than a dozen eggs, only one embryo was viable. In the next round of IVF, Janelle was not given enough local anaesthetic during the egg retrieval; by the end of the procedure, she was bawling her eyes out on the operating table. Still, the fertilization didn’t take.
The rounds of IVF were damaging to both Janelle’s body and mental health, and so, with COVID-19 raging across Europe, she took a break. It was only a couple of years later that she started trying again, and this time, for whatever reason, as it was for my mom, Janelle’s fourth round of IVF was successful. In 2022, a couple of days before her 30th birthday, Janelle gave birth to her first child.
It felt like a miracle—but could Janelle and Joe’s experience have been better if some of those human errors had been removed from the equation?
When I asked Janelle about how she would have felt about using AI in the IVF process, she paused for a second. “It’s so much harder to answer now having kids because there’s a definite desperation when you really really want a child,” she said. If the AI speeds up the process, “I can so see why people would be on board,” she said. “Whenever you’ve decided that you want to have a child, waiting one extra month is too long.”
But where do you draw the line? She asked. Is it OK to use an AI to select the embryos most likely to lead to a pregnancy? Probably. Is it OK to use an AI to select the embryos most likely not to carry a genetic disease? Possibly. Is it OK to use an AI to select the embryos for inherent genetic qualities, like intelligence or gender? No, thank you.
“There are so many nuances around it already without adding an AI,” Janelle said.
When Omer Benjakob and his wife Michal learned that IVF might help them conceive a child without passing along a rare genetic disease one of them suffers from, they didn’t think twice. They went to the hospital in Tel Aviv, where they live. After one round of IVF in which the embryos containing the malicious genetic code were discarded, their doctor told them the good news: Five healthy embryos had been fertilized and frozen, ready to be implanted whenever they wanted over the next five years. With the push of a button, they could have a baby.
In Israel, Benjakob tells me, “it’s very socially acceptable to have not just IVF, but a whole range of medical interventions around genetics.” He suggested that the discovery of DNA itself was one of the first examples of employing artificial intelligence in the field of genetics. “Our first use cases of AI were probably genetic,” he says. “It makes sense, right? Like think about DNA: It’s one of the most complex [things out there], huge amounts of numbers and variables imaginable.”
In other words, to answer complex, big data questions, it makes sense to have big data tools.
In some cases, the analysis can be pushed a whole lot further. In polygenic risk scoring, medical practitioners use deep-learning algorithms that promise to screen not only for genetic disorders, but also traits like intelligence and height. In the US, a cottage industry of companies offering these deep-learning solutions has emerged with names like LifeView and MyOme. In other countries, like the UK, rating embryos on genomic characteristics is illegal.
He suggested that the discovery of DNA itself was one of the first examples of employing artificial intelligence in the field of genetics.
In 2024, The Guardian published an investigation using undercover videos filmed by the activist group HOPE not hate. The videos show how, using advanced machine learning, one company is offering polygenic risk scoring. Heliospect Genomics reportedly charged prospective IVF parents $50,000 to rank up to 100 embryos on “IQ and the other naughty traits that everybody wants.”
“Everyone can have all the children they want and they can have children that are basically disease-free, smart, healthy; it’s going to be great,” the company’s CEO said in a pitch call shared with The Guardian.
Is this, in fact, the Brave New World so feared when IVF was first introduced: a class system rigged from birth where the highest bidder gets the best genes?
A group of researchers at Monash University in Australia has been pondering the difficult questions. In February 2025, they published a paper in European Society of Human Reproduction and Embryology: “Ethics of Artificial Intelligence in Embryo Assessment: Mapping the Terrain.” The paper was one of the first of its kind to take the question of ethics in AI-assisted human reproduction into the academic sphere.
“AI in IVF could be awesome, or it could be terrifying,” Catherine Mills, one of the researchers, tells me from her home office in Melbourne. “So the goal was to try to figure out where it lies between those two extremes.”
Mills first began looking into AI in IVF in 2021. A couple of years later, she teamed up with the study’s lead author, Julian Koplin, and the other researchers to get a better sense of “where things were at,” both in Australia and abroad, and “how people were thinking about these technologies within the industry,” Mills explained.
Using a grant from Ferring, a global pharmaceutical company, Mills and her colleagues started speaking with industry insiders, physicians, embryologists, and data scientists. What they found was surprising. “The people who were using AI already tended to both see it as less problematic ethically and less in need of extra regulation and so on, but also less useful,” Mills explained. “So, in using it, they came to see its limitations.”
While the objectivity and speed of AI tools were often cited as positives, real questions remain about their effectiveness and transparency, as well as the impact on the medical workforce over time.
Between January 2020 and September 2022, another group of researchers in Australia and Sweden set out to do a test: Man vs. machine. Using a dataset of 1,066 patients at 14 fertility clinics in Australia and Europe (Denmark, Sweden, and the United Kingdom), embryos were randomly assigned to be graded by either a human embryologist or a deep-learning algorithm called Data Analysis Score (iDAScore), which analyzes time-lapse images to predict the possibility of a clinical pregnancy.
It was the first large-scale, randomized trial. The results, published in 2024 in the scientific journal Nature, were a mixed bag. When the embryos were chosen by embryologists, clinical pregnancies resulted in 48 percent of cases; the AI lagged slightly below, at 46 percent. In roughly two-thirds of cases, the AI and the embryologist chose the same embryo. However, as lead study author Christos Venetis wrote in The Conversation, the AI made its choice 10 times more quickly.
“AI in IVF could be awesome, or it could be terrifying.”
To Iman Hajirasouliha, a computational biomedicine researcher at the Weill School of Medicine of Cornell University, the gap between the gushing rhetoric of AI in IVF pioneers and its clinical application is to be expected.
“I think there is some hype around AI and automation” in the fertility space, Hajirasouliha says. “You really need to validate these models—you need to show they are generalizable. We see that, for example, our model works very well—almost perfectly—in our clinic. But if you want to test it on different clinics, it’s still good, but not as accurate.”
The inevitable improvement of the models brings about other, more existential questions, Julian Koplin, the lead author of the Monash study, says. “You don’t know exactly why it’s recommending what it is recommending, and that means it can be hard to scrutinize the reasons why it suggests one thing over another,” Koplin explains. “You can imagine cases where the developers behaved responsibly, the clinicians used it appropriately, and then it goes rogue and something goes wrong. It’s really hard to know where to put the blame.”
What if one of these deep-learning tools were to begin suggesting discarding female embryos because of an unexplained bias inherent in the algorithm (as one recent study on embryo grading found), Koplin asks. And even if the technology does improve enough to consistently outpredict human embryologists, without input bias, to what extent will the human decision continue to carry weight?
Even for the most accomplished of embryologists, Koplin says, it might be hard to question the model once it’s been proven to be better than humans at picking the right embryo: “It would feel almost arrogant to have this thing . . . saying, ‘You should go this way,’ and then say, ‘Look, I know that my squishy human meat brain isn’t as good as that, but I’ve got this vibe, so I’m going to go this other way.’”
Eventually, AI tools could lead to a “deskilling” of embryologists, the Monash researchers feared. Koplin compares it to a drone operator using AI tools for determining strike targets on a battlefield: Eventually, it’s easier to defer to the objectivity of the machine than to make the tough life-or-death decision. “Everyone says there’s always going to be a human in the loop,” Mills says. “But, I think that’s naive. The dream of many people is to have a fully automated ART [assisted reproductive technology] clinic.”
Eventually, it’s easier to defer to the objectivity of the machine than to make the tough life-or-death decision.
And what about the patients in all of this? In their study, more than 40 percent of doctors who used AI tools in the IVF process felt that they didn’t need to warn patients that AI had been involved. But to Janelle, the IVF mom in New Zealand, knowing feels paramount. “I think it’s really important [to have] informed consent that AI’s being used, particularly because it’s your medical data,” she says. “Obviously, as someone who doesn’t understand all of the details about how AI works, [I’d want] to understand the limitations of where [that data] is stored and who gets access to it in the future.”
Mills highlighted that despite these questions, the Brave New World of AI in IVF was still far from existing.
“The way in which technologies like this are implemented is very, very staggered and ad hoc,” Mills said. “It’s easy to get caught in dystopian thinking if you just focus on the way things work in the West, to put it bluntly. While it might be that one clinic in Boston uses AI, as a matter of course, there are thousands of other clinics around the world that are nowhere near using AI.”
But future growth feels inevitable. Companies such as IVF 2.0, Vitrolife, Life Whisperer, FertilAI, and Nova IVF promise to optimize embryo selection through AI tools trained on time-lapse videos. ExSeed Health and LensHooke use an AI tool called CASA to analyze sperm samples. MIM Fertility in Poland developed a platform, Folliscan, that uses deep learning to monitor follicles—fluid-filled sacs in the ovaries that contain eggs—and recommend trigger timing.
These tools are increasingly being combined into one-size-fits-all platforms. In New York, Conceivable Life Sciences combines AI large language models with advanced robotics. In September 2025, the company announced that it had raised $50 million USD in venture capital funding to “amplify its work” in the automated IVF space.
“There’s still a lot to do, so the work is not done,” Gilboa, at AIVF in Israel, says. “But I think we’re on a good path forward.” Hariton, in California, agrees: “We are still in the dial-up days of AI . . . but I do think we’re gonna get to the point where the AIs can do more and can do it well—and probably can do some of it autonomously.”
A couple of days after my mom told me about my unlikely birth, I went on a bike ride through the golden fields of Champagne. When I got to a picnic table in a vast field where I often stop for a snack, I took out my journal and began to write. “The million different versions of me that might have existed,” I titled my entry.
After my mom’s extraction, on election day 1992, doctors proceeded to fertilize seven eligible ova. As my parents waited to see what the result of the procedure would be—the actual procedure that worked for them was a combination of IVF and GIFT (gamete intrafallopian transfer)—my dad took to calling the embryos the seven dwarves. By the time my mom went for her follow-up, I was the only embryo that survived. A couple of weeks later, the tone of my mom’s journal entries shifted markedly. “Well, our life is going to change!” she wrote cheerily.
A few days later, she had her first look at me in a sonogram, and then a month later, her next. “The first time I saw it, it looked like a planet in a universe—a clear circle floating in the uterus, w/ very little detail, just a nucleus inside,” she wrote. “This second time the image was bigger & much more complicated. AND . . . we saw the heartbeat! Amazing! It takes you to the awareness of the miracle at the center of life. No matter how well we understand it, how close we get to that moment of creation, it is still ineffable & ungraspable—as ungraspable as—on the other end—the fact of the death of a loved one.”
Sitting in the field, I thought back to my mom’s fear—several years before my birth—of “pushing the river,” and how she overcame that, how scary it must have been to try this relatively new technology for the first time. Later, speaking with Julian Koplin, the AI-sceptic researcher at Monash, I was struck by something he said: “I’d want [AI] to be used if my partner and I were going through IVF. I’d want anything that reduces the burdens and the costs of infertility.”
The current is in motion. Moving further along the river feels inevitable—but I do still wonder what happens when that river opens up into the ocean.
Phineas Rueckert is an independent journalist based in Paris. His writing has appeared in The Guardian, New Lines Magazine, The Nation, The Dial, Vice World News, Atlas Obscura, and elsewhere.
Editor: Carolyn Wells
Fact-checker: Julie Schwietert Collazo
Copyeditor: Krista Stevens
from Longreads https://longreads.com/2026/03/03/ivf-and-ai-infertility/
via IFTTT
Watch