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.
Recent leaps in technology represent an important step forward in unlocking artificial photosynthesis.
Since the beginning of life on Earth, plants have been naturally converting sunlight into energy. This photosynthesis process that's effortless for them has been anything but for scientists who have been trying to achieve artificial photosynthesis for the last half a century with the goal of creating a carbon-neutral fuel. Such a fuel could be a gamechanger — rather than putting CO2 back into the atmosphere like traditional fuels do, it would take CO2 out of the atmosphere and convert it into usable energy.
If given the option between a carbon-neutral fuel at the gas station and a fuel that produces carbon dioxide in spades -- and if costs and effectiveness were equal --who wouldn't choose the one best for the planet? That's the endgame scientists are after. A consumer switch to clean fuel could have a huge impact on our global CO2 emissions.
Up until this point, the methods used to make liquid fuel from atmospheric CO2 have been expensive, not efficient enough to really get off the ground, and often resulted in unwanted byproducts. But now, a new technology may be the key to unlocking the full potential of artificial photosynthesis. At the very least, it's a step forward and could help make a dent in atmospheric CO2 reduction.
"It's an important breakthrough in artificial photosynthesis," says Qian Wang, a researcher in the Department of Chemistry at Cambridge University and lead author on a recent study published in Nature about an innovation she calls "photosheets."
The latest version of the artificial leaf directly produces liquid fuel, which is easier to transport and use commercially.
These photosheets convert CO2, sunlight, and water into a carbon-neutral liquid fuel called formic acid without the aid of electricity. They're made of semiconductor powders that absorb sunlight. When in the presence of water and CO2, the electrons in the powders become excited and join with the CO2 and protons from the water molecules, reducing the CO2 in the process. The chemical reaction results in the production of formic acid, which can be used directly or converted to hydrogen, another clean energy fuel.
In the past, it's been difficult to reduce CO2 without creating a lot of unwanted byproducts. According to Wang, this new conversion process achieves the reduction and fuel creation with almost no byproducts.
The Cambridge team's new technology is a first and certainly momentous, but they're far from the only team to have produced fuel from CO2 using some form of artificial photosynthesis. More and more scientists are aiming to perfect the method in hopes of producing a truly sustainable, photosynthetic fuel capable of lowering carbon emissions.
Thanks to advancements in nanoscience, which has led to better control of materials, more successes are emerging. A team at the University of Illinois at Urbana-Champaign, for example, used gold nanoparticles as the photocatalysts in their process.
"My group demonstrated that you could actually use gold nanoparticles both as a light absorber and a catalyst in the process of converting carbon dioxide to hydrocarbons such as methane, ethane and propane fuels," says professor Prashant Jain, co-author of the study. Not only are gold nanoparticles great at absorbing light, they don't degrade as quickly as other metals, which makes them more sustainable.
That said, Jain's team, like every other research team working on artificial photosynthesis including the Cambridge team, is grappling with efficiency issues. Jain says that all parts of the process need to be optimized so the reaction can happen as quickly as possible.
"You can't just improve one [aspect], because that can lead to a decrease in performance in some other aspects," Jain explains.
The Cambridge team is currently experimenting with a range of catalysts to improve their device's stability and efficiency. Virgil Andrei, who is working on an artificial leaf design that was developed at Cambridge in 2019, was recently able to improve the performance and selectivity of the device. Now the leaf's solar-to-CO2 energy conversion efficiency is 0.2%, twice its previous efficiency.
The latest version also directly produces liquid fuel, which is easier to transport and use commercially.
In determining a method of fuel production's efficiency, one must consider how sustainable it is at every stage. That involves calculating whenever excess energy is needed to complete a step. According to Jain, in order to use CO2 for fuel production, you have to condense the CO2, which takes energy. And on the fuel production side, once the chemical reaction has created your byproducts, they need to be separated, which also takes energy.
To be truly sustainable, each part of the conversion system also needs to be durable. If parts need to be replaced often, or regularly maintained, that counts against it. Then you have to account for the system's reuse cycle. If you extract CO2 from the environment and convert it into fuel that's then put into a fuel cell, it's going to release CO2 at the other end. In order to create a fully green, carbon-neutral fuel source, that same amount of CO2 needs to be trapped and reintroduced back into the fuel conversion system.
"The cycle continues, and at each point, you will see a loss in efficiency, and depending on how much you [may also] see a loss in yield," says Jain. "And depending on what those efficiencies are at each one of those points will determine whether or not this process can be sustainable."
The science is at least a decade away from offering a competitive sustainable fuel option at scale. Streamlining a process to mimic what plants have perfected over billions of years is no small feat, but an ever-growing community of researchers using rapidly advancing technology is driving progress forward.
Genetic data sets skew too European, threatening to narrow who will benefit from future advances.
Genomics has begun its golden age. Just 20 years ago, sequencing a single genome cost nearly $3 billion and took over a decade. Today, the same feat can be achieved for a few hundred dollars and the better part of a day . Suddenly, the prospect of sequencing not just individuals, but whole populations, has become feasible.
The genetic differences between humans may seem meager, only around 0.1 percent of the genome on average, but this variation can have profound effects on an individual's risk of disease, responsiveness to medication, and even the dosage level that would work best.
Already, initiatives like the U.K.'s 100,000 Genomes Project - now expanding to 1 million genomes - and other similarly massive sequencing projects in Iceland and the U.S., have begun collecting population-scale data in order to capture and study this variation.
The resulting data sets are immensely valuable to researchers and drug developers working to design new 'precision' medicines and diagnostics, and to gain insights that may benefit patients. Yet, because the majority of this data comes from developed countries with well-established scientific and medical infrastructure, the data collected so far is heavily biased towards Western populations with largely European ancestry.
This presents a startling and fast-emerging problem: groups that are under-represented in these datasets are likely to benefit less from the new wave of therapeutics, diagnostics, and insights, simply because they were tailored for the genetic profiles of people with European ancestry.
We may indeed be approaching a golden age of genomics-enabled precision medicine. But if the data bias persists then there is a risk, as with most golden ages throughout history, that the benefits will not be equally accessible to all, and existing inequalities will only be exacerbated.
To remedy the situation, a number of initiatives have sprung up to sequence genomes of under-represented groups, adding them to the datasets and ensuring that they too will benefit from the rapidly unfolding genomic revolution.
Global Gene Corp
The idea behind Global Gene Corp was born eight years ago in Harvard when Sumit Jamuar, co-founder and CEO, met up with his two other co-founders, both experienced geneticists, for a coffee.
"They were discussing the limitless applications of understanding your genetic code," said Jamuar, a business executive from New Delhi.
"And so, being a technology enthusiast type, I was excited and I turned to them and said hey, this is incredible! Could you sequence me and give me some insights? And they actually just turned around and said no, because it's not going to be useful for you - there's not enough reference for what a good Sumit looks like."
What started as a curiosity-driven conversation on the power of genomics ended with a commitment to tackle one of the field's biggest roadblocks - its lack of global representation.
Jamuar set out to begin with India, which has about 20 percent of the world's population, including over 4000 different ethnicities, but contributes less than 2 percent of genomic data, he told Leaps.org.
Eight years later, Global Gene Corp's sequencing initiative is well underway, and is the largest in the history of the Indian subcontinent. The program is being carried out in collaboration with biotech giant Regeneron, with support from the Indian government, local communities, and the Indian healthcare ecosystem. In August 2020, Global Gene Corp's work was recognized through the $1 million 2020 Roddenberry award for organizations that advance the vision of 'Star Trek' creator Gene Roddenberry to better humanity.
This problem has already begun to manifest itself in, for example, much higher levels of genetic misdiagnosis among non-Europeans tested for their risk of certain diseases, such as hypertrophic cardiomyopathy - an inherited disease of the heart muscle.
Global Gene Corp also focuses on developing and implementing AI and machine learning tools to make sense of the deluge of genomic data. These tools are increasingly used by both industry and academia to guide future research by identifying particularly promising or clinically interesting genetic variants. But if the underlying data is skewed European, then the effectiveness of the computational analysis - along with the future advances and avenues of research that emerge from it - will be skewed towards Europeans too.
This problem has already begun to manifest itself in, for example, much higher levels of genetic misdiagnosis among non-Europeans tested for their risk of certain diseases, such as hypertrophic cardiomyopathy - an inherited disease of the heart muscle. Most of the genetic variants used in these tests were identified as being causal for the disease from studies of European genomes. However, many of these variants differ both in their distribution and clinical significance across populations, leading to many patients of non-European ancestry receiving false-positive test results - as their benign genetic variants were misclassified as pathogenic. Had even a small number of genomes from other ethnicities been included in the initial studies, these misdiagnoses could have been avoided.
"Unless we have a data set which is unbiased and representative, we're never going to achieve the success that we want," Jamuar says.
"When Siri was first launched, she could hardly recognize an accent which was not of a certain type, so if I was trying to speak to Siri, I would have to repeat myself multiple times and try to mimic an accent which wasn't my accent so that she could understand it.
"But over time the voice recognition technology improved tremendously because the training data was expanded to include people of very diverse backgrounds and their accents, so the algorithms were trained to be able to pick that up and it dramatically improved the technology. That's the way we have to think about it - without that good-quality diverse data, we will never be able to achieve the full potential of the computational tools."
While mapping India's rich genetic diversity has been the organization's primary focus so far, they plan, in time, to expand their work to other under-represented groups in Asia, the Middle East, Africa, and Latin America.
"As other like-minded people and partners join the mission, it just accelerates the achievement of what we have set out to do, which is to map out and organize the world's genomic diversity so that we can enable high-quality life and longevity benefits for everyone, everywhere," Jamuar says.
Empowering African Genomics
Africa is the birthplace of our species, and today still retains an inordinate amount of total human genetic diversity. Groups that left Africa and went on to populate the rest of the world, some 50 to 100,000 years ago, were likely small in number and only took a fraction of the total genetic diversity with them. This ancient bottleneck means that no other group in the world can match the level of genetic diversity seen in modern African populations.
Despite Africa's central importance in understanding the history and extent of human genetic diversity, the genomics of African populations remains wildly understudied. Addressing this disparity has become a central focus of the H3Africa Consortium, an initiative formally launched in 2012 with support from the African Academy of Sciences, the U.S. National Institutes of Health, and the UK's Wellcome Trust. Today, H3Africa supports over 50 projects across the continent, on an array of different research areas in genetics relevant to the health and heredity of Africans.
"Africa is the cradle of Humankind. So what that really means is that the populations that are currently living in Africa are among some of the oldest populations on the globe, and we know that the longer populations have had to go through evolutionary phases, the more variation there is in the genomes of people who live presently," says Zane Lombard, a principal investigator at H3Africa and Associate Professor of Human Genetics at the University of the Witwatersrand in Johannesburg, South Africa.
"So for that reason, African populations carry a huge amount of genetic variation and diversity, which is pretty much uncaptured. There's still a lot to learn as far as novel variation is concerned by looking at and studying African genomes."
A recent landmark H3Africa study, led by Lombard and published in Nature in October, sequenced the genomes of over 400 African individuals from 50 ethno-linguistic groups - many of which had never been sampled before.
Despite the relatively modest number of individuals sequenced in the study, over three million previously undescribed genetic variants were found, and complex patterns of ancestral migration were uncovered.
"In some of these ethno-linguistic groups they don't have a word for DNA, so we've had to really think about how to make sure that we communicate the purposes of different studies to participants so that you have true informed consent," says Lombard.
"The objective," she explained, "was to try and fill some of the gaps for many of these populations for which we didn't have any whole genome sequences or any genetic variation data...because if we're thinking about the future of precision medicine, if the patient is a member of a specific group where we don't know a lot about the genomic variation that exists in that group, it makes it really difficult to start thinking about clinical interpretation of their data."
From H3Africa's conception, the consortium's goal has not only been to better represent Africa's staggering genetic diversity in genomic data sets, but also to build Africa's domestic genomics capabilities and empower a new generation of African researchers. By doing so, the hope is that Africans will be able to set their own genomics agenda, and leapfrog to new and better ways of doing the work.
"The training that has happened on the continent and the number of new scientists, new students, and fellows that have come through the process and are now enabled to start their own research groups, to grow their own research in their countries, to be a spokesperson for genomics research in their countries, and to build that political will to do these larger types of sequencing initiatives - that is really a significant outcome from H3Africa as well. Over and above all the science that's coming out," Lombard says.
"What has been created through H3Africa is just this locus of researchers and scientists and bioethicists who have the same goal at heart - to work towards adjusting the data bias and making sure that all global populations are represented in genomics."