Your Digital Avatar May One Day Get Sick Before You Do
Artificial neurons in a concept of artificial intelligence.
Artificial intelligence is everywhere, just not in the way you think it is.
These networks, loosely designed after the human brain, are interconnected computers that have the ability to "learn."
"There's the perception of AI in the glossy magazines," says Anders Kofod-Petersen, a professor of Artificial Intelligence at the Norwegian University of Science and Technology. "That's the sci-fi version. It resembles the small guy in the movie AI. It might be benevolent or it might be evil, but it's generally intelligent and conscious."
"And this is, of course, as far from the truth as you can possibly get."
What Exactly Is Artificial Intelligence, Anyway?
Let's start with how you got to this piece. You likely came to it through social media. Your Facebook account, Twitter feed, or perhaps a Google search. AI influences all of those things, machine learning helping to run the algorithms that decide what you see, when, and where. AI isn't the little humanoid figure; it's the system that controls the figure.
"AI is being confused with robotics," Eleonore Pauwels, Director of the Anticipatory Intelligence Lab with the Science and Technology Innovation Program at the Wilson Center, says. "What AI is right now is a data optimization system, a very powerful data optimization system."
The revolution in recent years hasn't come from the method scientists and other researchers use. The general ideas and philosophies have been around since the late 1960s. Instead, the big change has been the dramatic increase in computing power, primarily due to the development of neural networks. These networks, loosely designed after the human brain, are interconnected computers that have the ability to "learn." An AI, for example, can be taught to spot a picture of a cat by looking at hundreds of thousands of pictures that have been labeled "cat" and "learning" what a cat looks like. Or an AI can beat a human at Go, an achievement that just five years ago Kofod-Petersen thought wouldn't be accomplished for decades.
"It's very difficult to argue that something is intelligent if it can't learn, and these algorithms are getting pretty good at learning stuff. What they are not good at is learning how to learn."
Medicine is the field where this expertise in perception tasks might have the most influence. It's already having an impact as iPhones use AI to detect cancer, Apple watches alert the wearer to a heart problem, AI spots tuberculosis and the spread of breast cancer with a higher accuracy than human doctors, and more. Every few months, another study demonstrates more possibility. (The New Yorker published an article about medicine and AI last year, so you know it's a serious topic.)
But this is only the beginning. "I personally think genomics and precision medicine is where AI is going to be the biggest game-changer," Pauwels says. "It's going to completely change how we think about health, our genomes, and how we think about our relationship between our genotype and phenotype."
The Fundamental Breakthrough That Must Be Solved
To get there, however, researchers will need to make another breakthrough, and there's debate about how long that will take. Kofod-Petersen explains: "If we want to move from this narrow intelligence to this broader intelligence, that's a very difficult problem. It basically boils down to that we haven't got a clue about what intelligence actually is. We don't know what intelligence means in a biological sense. We think we might recognize it but we're not completely sure. There isn't a working definition. We kind of agree with the biologists that learning is an aspect of it. It's very difficult to argue that something is intelligent if it can't learn, and these algorithms are getting pretty good at learning stuff. What they are not good at is learning how to learn. They can learn specific tasks but we haven't approached how to teach them to learn to learn."
In other words, current AI is very, very good at identifying that a picture of a cat is, in fact, a cat – and getting better at doing so at an incredibly rapid pace – but the system only knows what a "cat" is because that's what a programmer told it a furry thing with whiskers and two pointy ears is called. If the programmer instead decided to label the training images as "dogs," the AI wouldn't say "no, that's a cat." Instead, it would simply call a furry thing with whiskers and two pointy ears a dog. AI systems lack the explicit inference that humans do effortlessly, almost without thinking.
Pauwels believes that the next step is for AI to transition from supervised to unsupervised learning. The latter means that the AI isn't answering questions that a programmer asks it ("Is this a cat?"). Instead, it's almost like it's looking at the data it has, coming up with its own questions and hypothesis, and answering them or putting them to the test. Combining this ability with the frankly insane processing power of the computer system could result in game-changing discoveries.
In the not-too-distant future, a doctor could run diagnostics on a digital avatar, watching which medical conditions present themselves before the person gets sick in real life.
One company in China plans to develop a way to create a digital avatar of an individual person, then simulate that person's health and medical information into the future. In the not-too-distant future, a doctor could run diagnostics on a digital avatar, watching which medical conditions presented themselves – cancer or a heart condition or anything, really – and help the real-life version prevent those conditions from beginning or treating them before they became a life-threatening issue.
That, obviously, would be an incredibly powerful technology, and it's just one of the many possibilities that unsupervised AI presents. It's also terrifying in the potential for misuse. Even the term "unsupervised AI" brings to mind a dystopian landscape where AI takes over and enslaves humanity. (Pick your favorite movie. There are dozens.) This is a concern, something for developers, programmers, and scientists to consider as they build the systems of the future.
The Ethical Problem That Deserves More Attention
But the more immediate concern about AI is much more mundane. We think of AI as an unbiased system. That's incorrect. Algorithms, after all, are designed by someone or a team, and those people have explicit or implicit biases. Intentionally, or more likely not, they introduce these biases into the very code that forms the basis for the AI. Current systems have a bias against people of color. Facebook tried to rectify the situation and failed. These are two small examples of a larger, potentially systemic problem.
It's vital and necessary for the people developing AI today to be aware of these issues. And, yes, avoid sending us to the brink of a James Cameron movie. But AI is too powerful a tool to ignore. Today, it's identifying cats and on the verge of detecting cancer. In not too many tomorrows, it will be on the forefront of medical innovation. If we are careful, aware, and smart, it will help simulate results, create designer drugs, and revolutionize individualize medicine. "AI is the only way to get there," Pauwels says.
Researchers Are Discovering How to Predict – and Maybe Treat — Pregnancy Complications Early On.
Katie Love cradles her newborn daughter, born after a bout with preeclampsia.
Katie Love wishes there was some way she could have been prepared. But there was no way to know, early in 2020, that her pregnancy would lead to terrifyingly high blood pressure and multiple hospital visits, ending in induced labor and a 56-hour-long, “nightmare” delivery at 37 weeks. Love, a social media strategist in Pittsburgh, had preeclampsia, a poorly understood and potentially deadly pregnancy complication that affects 1 in 25 pregnant women in the United States. But there was no blood test, no easy diagnostic marker to warn Love that this might happen. Even on her first visit to the emergency room, with sky-high blood pressure, doctors could not be certain preeclampsia was the cause.
In fact, the primary but imperfect indicators for preeclampsia — high blood pressure and protein in the urine — haven’t changed in decades. The Preeclampsia Foundation calls a simple, rapid test to predict or diagnose the condition “a key component needed in the fight.”
Another common pregnancy complication is preterm birth, which affects 1 in 10 U.S. pregnancies, but there are few options to predict that might happen, either.
“The best tool that obstetricians have at the moment is still a tape measure and a blood pressure cuff to diagnose whatever’s happening in your pregnancy,” says Fiona Kaper, a vice president at the DNA-sequencing company Illumina in San Diego.
The hunt for such specific biomarkers is now taking off, at Illumina and elsewhere, as scientists probe maternal blood for signs that could herald pregnancy problems. These same molecules offer clues that might lead to more specific treatments. So far, it’s clear that many complications start with the placenta, the temporary organ that transfers nutrients, oxygen and waste between mother and fetus, and that these problems often start well before symptoms arise. Researchers are using the latest stem-cell technology to better understand the causes of complications and test treatments.
Pressing Need
Obstetricians aren’t flying completely blind; medical history can point to high or low risk for pregnancy complications. But ultimately, “everybody who’s pregnant is at risk for preeclampsia,” says Sarosh Rana, chief of maternal-fetal medicine at University of Chicago Medicine and an advisor to the Preeclampsia Foundation. And the symptoms of the condition include problems like headache and swollen feet that overlap with those of pregnancy in general, complicating diagnoses.
The “holy grail" would be early, first-trimester biomarkers. If obstetricians and expecting parents could know, in the first few months of pregnancy, that preeclampsia is a risk, a pregnant woman could monitor her blood pressure at home and take-low dose aspirin that might stave it off.
There are a couple more direct tests physicians can turn to, but these are imperfect. For preterm labor, fetal fibronectin makes up a sort of glue that keeps the amniotic sac, which cushions the unborn baby, attached to the uterus. If it’s not present near a woman’s cervix, that’s a good indicator that she’s not in labor, and can be safely sent home, says Lauren Demosthenes, an obstetrician and senior medical director of the digital health company Babyscripts in Washington, D.C. But if fibronectin appears, it might or might not indicate preterm labor.
“What we want is a test that gives us a positive predictive [signal],” says Demosthenes. “I want to know, if I get it, is it really going to predict preterm birth, or is it just going to make us worry more and order more tests?” In fact, the fetal fibronectin test hasn’t been shown to improve pregnancy outcomes, and Demosthenes says it’s fallen out of favor in many clinics.
Similarly, there’s a blood test, based on the ratio of the amounts of two different proteins, that can rule out preeclampsia but not confirm it’s happening. It’s approved in many countries, though not the U.S.; studies are still ongoing. A positive test, which means “maybe preeclampsia,” still leaves doctors and parents-to-be facing excruciating decisions: If the mother’s life is in danger, delivering the baby can save her, but even a few more days in the uterus can promote the baby’s health. In Ireland, where the test is available, it’s not getting much use, says Patricia Maguire, director of the University College Dublin Institute for Discovery.
Maguire has identified proteins released by platelets that indicate pregnancy — the “most expensive pregnancy test in the world,” she jokes. She is now testing those markers in women with suspected preeclampsia.
The “holy grail,” says Maguire, would be early, first-trimester biomarkers. If obstetricians and expecting parents could know, in the first few months of pregnancy, that preeclampsia is a risk, a pregnant woman could monitor her blood pressure at home and take-low dose aspirin that might stave it off. Similarly, if a quick blood test indicated that preterm labor could happen, doctors could take further steps such as measuring the cervix and prescribing progesterone if it’s on the short side.
Biomarkers in Blood
It was fatherhood that drew Stephen Quake, a biophysicist at Stanford University in California, to the study of pregnancy biomarkers. His wife, pregnant with their first child in 2001, had a test called amniocentesis. That involves extracting a sample from within the uterus, using a 3–8-inch-long needle, for genetic testing. The test can identify genetic differences, such as Down syndrome, but also carries risks including miscarriage or infection. In this case, mom and baby were fine (Quake’s daughter is now a college student), but he found the diagnostic danger unacceptable.
Seeking a less invasive test, Quake in 2008 reported that there’s enough fetal DNA in the maternal bloodstream to diagnose Down syndrome and other genetic conditions. “Use of amniocentesis has plunged,” he says.
Then, recalling that his daughter was born three and a half weeks before her due date — and that Quake’s own mom claims he was a month late, which makes him think the due date must have been off — he started researching markers that could accurately assess a fetus’ age and predict the timing of labor. In this case, Quake was interested in RNA, not DNA, because it’s a signal of which genes the fetus’, placenta’s, and mother’s tissues are using to create proteins. Specifically, these are RNAs that have exited the cells that made them. Tissues can use such free RNAs as messages, wrapping them in membranous envelopes to travel the bloodstream to other body parts. Dying cells also release fragments containing RNAs. “A lot of information is in there,” says Kaper.
In a small study of 31 healthy pregnant women, published in 2018, Quake and collaborators discovered nine RNAs that could predict gestational age, which indicates due date, just as well as ultrasound. With another set of 38 women, including 13 who delivered early, the researchers discovered seven RNAs that predicted preterm labor up to two months in advance.
Quake notes that an RNA-based blood test is cheaper and more portable than ultrasound, so it might be useful in the developing world. A company he cofounded, Mirvie, Inc., is now analyzing RNA’s predictive value further, in thousands of diverse women. CEO and cofounder Maneesh Jain says that since preterm labor is so poorly understood, they’re sequencing RNAs that represent about 20,000 genes — essentially all the genes humans have — to find the very best biomarkers. “We don’t know enough about this field to guess what it might be,” he says. “We feel we’ve got to cast the net wide.”
Quake, and Mirvie, are now working on biomarkers for preeclampsia. In a recent preprint study, not yet reviewed by other experts, Quake’s Stanford team reported 18 RNAs that, measured before 16 weeks, correctly predicted preeclampsia 56–100% of the time.
Other researchers are taking a similar tack. Kaper’s team at Illumina was able to classify preeclampsia from bloodstream RNAs with 85 to 89% accuracy, though they didn’t attempt to predict it. And Louise Laurent, a maternal-fetal medicine specialist and researcher at the University of California, San Diego (UCSD), has defined several pairs of microRNAs — pint-sized RNAs that regulate other ones — in second-trimester blood samples that predict preeclampsia later on.
Placentas in a Dish
The RNAs that show up in these studies often come from genes used by the placenta. But they’re only signals that something’s wrong, not necessarily the root cause. “There still is not much known about what really causes major complications of pregnancy,” says Laurent.
The challenge is that placental problems likely occur early on, as the organ forms in the first trimester. For example, if the placenta did a poor job of building blood vessels through the uterine lining, it might cause preeclampsia later as the growing fetus tries to access more and more blood through insufficient vessels, leading to high blood pressure in the mother. “Everyone has kind of suspected that that is probably what goes wrong,” says Mana Parast, a pathologist and researcher at UCSD.
To see how a placenta first faltered, “you want to go back in time,” says Parast. It’s only recently become possible to do something akin to that: She and Laurent take cells from the umbilical cord (which is a genetic match for the placenta) at the end of pregnancy, and turn them into stem cells, which can become any kind of cell. They then nudge those stem cells to make new placenta cells in lab dishes. But when the researchers start with cells from an umbilical cord after preeclampsia, they find the stem cells struggle to even form proper placenta cells, or they develop abnormally. So yes, something seems to go wrong right at the beginning. Now, the team plans to use these cell cultures to study the microRNAs that indicate preeclampsia risk, and to look for medications that might reverse the problems, Parast says.
Biomarkers could lead to treatments. For example, one of the proteins that commercial preeclampsia diagnostic kits test for is called soluble Flt-1. It’s a sort of anti-growth factor, explains Rana, that can cause problems with blood vessels and thus high blood pressure. Getting rid of the extra Flt-1, then, might alleviate symptoms and keep the mother safe, giving the baby more time to develop. Indeed, a small trial that filtered this protein from the blood did lower blood pressure, allowing participants to keep their babies inside for a couple of weeks longer, researchers reported in 2011.
For pregnant women like Love, even advance warning would have been beneficial. Laurent and others envision a first-trimester blood test that would use different kinds of biomolecules — RNAs, proteins, whatever works best — to indicate whether a pregnancy is at low, medium, or high risk for common complications.
“I prefer to be prepared,” says Love, now the mother of a healthy little girl. “I just wouldn’t have been so thrown off by the whole thing.”
Dec. 17th Event: The Latest on Omicron, Boosters, and Immunity
The Omicron variant poses new uncertainty for the vaccines, which four leading experts will address during our virtual event on December 17th, 2021.
This virtual event will convene leading scientific and medical experts to discuss the most pressing questions around the new Omicron variant, including what we know so far about its ability to evade COVID-19 vaccines, the role of boosters in eliciting heightened immunity, and the science behind variants and vaccines. A public Q&A will follow the expert discussion.
EVENT INFORMATION:
Date: Friday Dec 17, 2021
2:00pm - 3:30pm EST
Dr. Céline Gounder, MD, ScM, is the CEO/President/Founder of Just Human Productions, a non-profit multimedia organization. She is also the host and producer of American Diagnosis, a podcast on health and social justice, and Epidemic, a podcast about infectious disease epidemics and pandemics. She served on the Biden-Harris Transition COVID-19 Advisory Board.
Dr. Theodora Hatziioannou, Ph.D., is a Research Associate Professor in the Laboratory of Retrovirology at The Rockefeller University. Her research includes identifying plasma samples from recovered COVID-19 patients that contain antibodies capable of neutralizing the SARS-CoV-2 coronavirus.
Dr. Onyema Ogbuagu, MBBCh, is an Associate Professor at Yale School of Medicine and an infectious disease specialist who treats COVID-19 patients and leads Yale’s clinical studies around COVID-19. He ran Yale’s trial of the Pfizer/BioNTech vaccine.
Dr. Eric Topol, M.D., is a cardiologist, scientist, professor of molecular medicine, and the director and founder of Scripps Research Translational Institute. He has led clinical trials in over 40 countries with over 200,000 patients and pioneered the development of many routinely used medications.
This event is the fourth of a four-part series co-hosted by Leaps.org, the Aspen Institute Science & Society Program, and the Sabin–Aspen Vaccine Science & Policy Group, with generous support from the Gordon and Betty Moore Foundation and the Howard Hughes Medical Institute.
Kira Peikoff was the editor-in-chief of Leaps.org from 2017 to 2021. As a journalist, her work has appeared in The New York Times, Newsweek, Nautilus, Popular Mechanics, The New York Academy of Sciences, and other outlets. She is also the author of four suspense novels that explore controversial issues arising from scientific innovation: Living Proof, No Time to Die, Die Again Tomorrow, and Mother Knows Best. Peikoff holds a B.A. in Journalism from New York University and an M.S. in Bioethics from Columbia University. She lives in New Jersey with her husband and two young sons. Follow her on Twitter @KiraPeikoff.