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.
In early 2020, Moderna Inc. was a barely-known biotechnology company with an unproven approach. It wanted to produce messenger RNA molecules to carry instructions into the body, teaching it to ward off disease. Experts doubted the Boston-based company would meet success.
Today, Moderna is a pharmaceutical power thanks to its success developing an effective Covid-19 vaccine. The company is worth $124 billion, more than giants including GlaxoSmithKline and Sanofi, and evidence has emerged that Moderna's shots are more protective than those produced by Pfizer-BioNTech and other vaccine makers. Pressure is building on the company to deliver more of its doses to people around the world, especially in poorer countries, and Moderna is working on vaccines against other pathogens, including Zika, influenza and cytomegalovirus.
But Moderna encountered such difficulties over the course of its eleven-year history that some executives worried it wouldn't survive. Two unlikely scientists helped save the company. Their breakthroughs paved the way for Moderna's Covid-19 shots but their work has never been publicized nor have their contributions been properly appreciated.
Derrick Rossi, a scientist at MIT, and Noubar Afeyan, a Cambridge-based investor, launched Moderna in September 2010. Their idea was to create mRNA molecules capable of delivering instructions to the body's cells, directing them to make proteins to heal ailments and cure disease. Need a statin, immunosuppressive, or other drug or vaccine? Just use mRNA to send a message to the body's cells to produce it. Rossi and Afeyan were convinced injecting mRNA into the body could turn it into its own laboratory, generating specific medications or vaccines as needed.
At the time, the notion that one might be able to teach the body to make proteins bordered on heresy. Everyone knew mRNA was unstable and set off the body's immune system on its way into cells. But in the late 2000's, two scientists at the University of Pennsylvania, Katalin Karikó and Drew Weissman, had figured out how to modify mRNA's chemical building blocks so the molecule could escape the notice of the immune system and enter the cell. Rossi and Afeyan couldn't convince the University of Pennsylvania to license Karikó and Weissman's patent, however, stymying Moderna's early ambitions. At the same time, the Penn scientists' technique seemed more applicable to an academic lab than a biotech company that needed to produce drugs or shots consistently and in bulk. Rossi and Afeyan's new company needed their own solution to help mRNA evade the body's defenses.
Some of Moderna's founders doubted Schrum could find success and they worried if their venture was doomed from the start.
The Scientist Who Modified mRNA: Jason Schrum
In 2010, Afeyan's firm subleased laboratory space in the basement of another Cambridge biotech company to begin scientific work. Afeyan chose a young scientist on his staff, Jason Schrum, to be Moderna's first employee, charging him with getting mRNA into cells without relying on Karikó and Weissman's solutions.
Schrum seemed well suited for the task. Months earlier, he had received a PhD in biological chemistry at Harvard University, where he had focused on nucleotide chemistry. Schrum even had the look of someone who might do big things. The baby-faced twenty-eight-year-old favored a relaxed, start-up look: khakis, button-downs, and Converse All-Stars.
Schrum felt immediate strain, however. He hadn't told anyone, but he was dealing with intense pain in his hands and joints, a condition that later would be diagnosed as degenerative arthritis. Soon Schrum couldn't bend two fingers on his left hand, making lab work difficult. He joined a drug trial, but the medicine proved useless. Schrum tried corticosteroid injections and anti-inflammatory drugs, but his left hand ached, restricting his experiments.
"It just wasn't useful," Schrum says, referring to his tender hand.
He persisted, nonetheless. Each day in the fall of 2010, Schrum walked through double air-locked doors into a sterile "clean room" before entering a basement laboratory, in the bowels of an office in Cambridge's Kendall Square neighborhood, where he worked deep into the night. Schrum searched for potential modifications of mRNA nucleosides, hoping they might enable the molecule to produce proteins. Like all such rooms, there were no windows, so Schrum had to check a clock to know if it was day or night. A colleague came to visit once in a while, but most of the time, Schrum was alone.
Some of Moderna's founders doubted Schrum could find success and they worried if their venture was doomed from the start. An established MIT scientist turned down a job with the start-up to join pharmaceutical giant Novartis, dubious of Moderna's approach. Colleagues wondered if mRNA could produce proteins, at least on a consistent basis.
As Schrum began testing the modifications in January 2011, he made an unexpected discovery. Karikó and Weissman saw that by turned one of the building blocks for mRNA, a ribonucleoside called uridine, into a slightly different form called pseudouridine, the cell's immune system ignored the mRNA and the molecule avoided an immune response. After a series of experiments in the basement lab, Schrum discovered that a variant of pseudouridine called N1- methyl-pseudouridine did an even better job reducing the cell's innate immune response. Schrum's nucleoside switch enabled even higher protein production than Karikó and Weissman had generated, and Schrum's mRNAs lasted longer than either unmodified molecules or the modified mRNA the Penn academics had used, startling the young researcher. Working alone in a dreary basement and through intense pain, he had actually improved on the Penn professors' work.
Years later, Karikó and Weissman who would win acclaim. In September 2021, the scientists were awarded the Lasker-DeBakey Clinical Medical Research Award. Some predict they eventually will win a Nobel prize. But it would be Schrum's innovation that would form the backbone of both Moderna and Pfizer-BioNTech's Covid-19 vaccine, not the chemical modifications that Karikó and Weissman developed. For Schrum, necessity had truly been the mother of invention.
The Scientist Who Solved Delivery: Kerry Benenato
For several years, Moderna would make slow progress developing drugs to treat various diseases. Eventually, the company decided that mRNA was likely better suited for vaccines. By 2017, Moderna and the National Institutes of Health were discussing working together to develop mRNA–based vaccines, a partnership that buoyed Moderna's executives. There remained a huge obstacle in Moderna's way, however. It was up to Kerry Benenato to find a solution.
Benenato received an early hint of the hurdle in front of her three years earlier, when the organic chemist was first hired. When a colleague gave her a company tour, she was introduced to Moderna's chief scientific officer, Joseph Bolen, who seemed unusually excited to meet her.
"Oh, great!" Bolen said with a smile. "She's the one who's gonna solve delivery."
Bolen gave a hearty laugh and walked away, but Benenato detected seriousness in his quip.
It was a lot to expect from a 37-year-old scientist already dealing with insecurities and self-doubt. Benenato was an accomplished researcher who most recently had worked at AstraZeneca after completing post-doctoral studies at Harvard University. Despite her impressive credentials, Benenato battled a lack of confidence that sometimes got in her way. Performance reviews from past employers had been positive, but they usually produced similar critiques: Be more vocal. Do a better job advocating for your ideas. Give us more, Kerry.
Benenato was petite and soft-spoken. She sometimes stuttered or relied on "ums" and "ahs" when she became nervous, especially in front of groups, part of why she sometimes didn't feel comfortable speaking up.
"I'm an introvert," she says. "Self-confidence is something that's always been an issue."
To Benenato, Moderna's vaccine approach seemed promising—the team was packaging mRNAs in microscopic fatty-acid compounds called lipid nanoparticles, or LNPs, that protected the molecules on their way into cells. Moderna's shots should have been producing ample and long-lasting proteins. But the company's scientists were alarmed—they were injecting shots deep into the muscle of mice, but their immune systems were mounting spirited responses to the foreign components of the LNPs, which had been developed by a Canadian company.
This toxicity was a huge issue: A vaccine or drug that caused sharp pain and awful fevers wasn't going to prove very popular. The Moderna team was in a bind: Its mRNA had to be wrapped in the fatty nanoparticles to have a chance at producing plentiful proteins, but the body wasn't tolerating the microscopic encasements, especially upon repeated dosing.
The company's scientists had done everything they could to try to make the molecule's swathing material disappear soon after entering the cells, in order to avoid the unfortunate side effects, such as chills and headaches, but they weren't making headway. Frustration mounted. Somehow, the researchers had to find a way to get the encasements—made of little balls of fat, cholesterol, and other substances—to deliver their payload mRNA and then quickly vanish, like a parent dropping a teenager off at a party, to avoid setting off the immune system in unpleasant ways, even as the RNA and the proteins the molecule created stuck around.
Benenato wasn't entirely shocked by the challenges Moderna was facing. One of the reasons she had joined the upstart company was to help develop its delivery technology. She just didn't realize how pressing the issue was, or how stymied the researchers had become. Benenato also didn't know that Moderna board members were among those most discouraged by the delivery issue. In meetings, some of them pointed out that pharmaceutical giants like Roche Holding and Novartis had worked on similar issues and hadn't managed to develop lipid nanoparticles that were both effective and well tolerated by the body. Why would Moderna have any more luck?
Stephen Hoge insisted the company could yet find a solution.
"There's no way the only innovations in LNP are going to come from some academics and a small Canadian company," insisted Hoge, who had convinced the executives that hiring Benenato might help deliver an answer.
Benenato realized that while Moderna might have been a hot Boston-area start- up, it wasn't set up to do the chemistry necessary to solve their LNP problem. Much of its equipment was old or secondhand, and it was the kind used to tinker with mRNAs, not lipids.
"It was scary," she says.
When Benenato saw the company had a nuclear magnetic resonance spectrometer, which allows chemists to see the molecular structure of material, she let out a sigh of relief. Then Benenato inspected the machine and realized it was a jalopy. The hulking, aging instrument had been decommissioned and left behind by a previous tenant, too old and banged up to bring with them.
Benenato began experimenting with different chemical changes for Moderna's LNPs, but without a working spectrometer she and her colleagues had to have samples ready by noon each day, so they could be picked up by an outside company that would perform the necessary analysis. After a few weeks, her superiors received an enormous bill for the outsourced work and decided to pay to get the old spectrometer running again.
After months of futility, Benenato became impatient. An overachiever who could be hard on herself, she was eager to impress her new bosses. Benenato felt pressure outside the office, as well. She was married with a preschool-age daughter and an eighteen-month-old son. In her last job, Benenato's commute had been a twenty-minute trip to Astra-Zeneca's office in Waltham, outside Boston; now she was traveling an hour to Moderna's Cambridge offices. She became anxious—how was she going to devote the long hours she realized were necessary to solve their LNP quandary while providing her children proper care? Joining Moderna was beginning to feel like a possible mistake.
She turned to her husband and father for help. They reminded her of the hard work she had devoted to establishing her career and said it would be a shame if she couldn't take on the new challenge. Benenato's husband said he was happy to stay home with the kids, alleviating some of her concerns.
Back in the office, she got to work. She wanted to make lipids that were easier for the body to chop into smaller pieces, so they could be eliminated by the body's enzymes. Until then, Moderna, like most others, relied on all kinds of complicated chemicals to hold its LNP packaging together. They weren't natural, though, so the body was having a hard time breaking them down, causing the toxicity.
Benenato began experimenting with simpler chemicals. She inserted "ester bonds"—compounds referred to in chemical circles as "handles" because the body easily grabs them and breaks them apart. Ester bonds had two things going for them: They were strong enough to help ensure the LNP remained stable, acting much like a drop of oil in water, but they also gave the body's enzymes something to target and break down as soon as the LNP entered the cell, a way to quickly rid the body of the potentially toxic LNP components. Benenato thought the inclusion of these chemicals might speed the elimination of the LNP delivery material.
This idea, Benenato realized, was nothing more than traditional, medicinal chemistry. Most people didn't use ester bonds because they were pretty unsophisticated. But, hey, the tricky stuff wasn't working, so Benenato thought she'd see if the simple stuff worked.
Benenato also wanted to try to replace a group of unnatural chemicals in the LNP that was contributing to the spirited and unwelcome response from the immune system. Benenato set out to build a new and improved chemical combination. She began with ethanolamine, a colorless, natural chemical, an obvious start for any chemist hoping to build a more complex chemical combination. No one relied on ethanolamine on its own.
Benenato was curious, though. What would happen if she used just these two simple modifications to the LNP: ethanolamine with the ester bonds? Right away, Benenato noticed her new, super-simple compound helped mRNA create some protein in animals. It wasn't much, but it was a surprising and positive sign. Benenato spent over a year refining her solution, testing more than one hundred variations, all using ethanolamine and ester bonds, showing improvements with each new version of LNP. After finishing her 102nd version of the lipid molecule, which she named SM102, Benenato was confident enough in her work to show it to Hoge and others.
They immediately got excited. The team kept tweaking the composition of the lipid encasement. In 2017, they wrapped it around mRNA molecules and injected the new combination in mice and then monkeys. They saw plentiful, potent proteins were being produced and the lipids were quickly being eliminated, just as Benenato and her colleagues had hoped. Moderna had its special sauce.
That year, Benenato was asked to deliver a presentation to Stephane Bancel, Moderna's chief executive, Afeyan, and Moderna's executive committee to explain why it made sense to use the new, simpler LNP formulation for all its mRNA vaccines. She still needed approval from the executives to make the change. Ahead of the meeting, she was apprehensive, as some of her earlier anxieties returned. But an unusual calm came over her as she began speaking to the group. Benenato explained how experimenting with basic, overlooked chemicals had led to her discovery.
She said she had merely stumbled onto the company's solution, though her bosses understood the efforts that had been necessary for the breakthrough. The board complimented her work and agreed with the idea of switching to the new LNP. Benenato beamed with pride.
"As a scientist, serendipity has been my best friend," she told the executives.
Over the next few years, Benenato and her colleagues would improve on their methods and develop even more tolerable and potent LNP encasement for mRNA molecules. Their work enabled Moderna to include higher doses of vaccine in its shots. In early 2020, Moderna developed Covid-19 shots that included 100 micrograms of vaccine, compared with 30 micrograms in the Pfizer-BioNTech vaccine. That difference appears to help the Moderna vaccine generate higher titers and provide more protection.
"You set out in a career in drug discovery to want to make a difference," Benenato says. "Seeing it come to reality has been surreal and emotional."
Editor's Note: This essay is excerpted from A SHOT TO SAVE THE WORLD: The Inside Story of the Life-or-Death Race for a COVID-19 Vaccine by Gregory Zuckerman, now on sale from Portfolio/Penguin.
In just 100 years, scientific breakthroughs could completely transform humanity and our planet for the better. Here's a glimpse at what our future may hold.
The Next 100 Years of Scientific Progress
Kira Peikoff is the editor-in-chief of Leaps.org. 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.