I shrug and put the phone back in my pocket. “No, it’s not.”
After a pause, she demands, “Let me see that again.”
The second time around, her eyes travel across the screen, taking time to study the image. She notes my angular brows. The sheen of silver on my breastplate. The slight raise of my chin. The effect of seeing your parents outside their usual circumstances is a little like seeing them in the nude. Embarrassing for everyone.
“Mom!” she cries at last, her voice equal parts wondering and dismayed. “It looks like you, but it isn’t you.”
Well, she’s right. It’s not the version of myself I display to her. The version she sees is usually in leggings with a stray hole along the seam, wearing no makeup, in a rush to pack a peanut-free snack while practicing a Vietnamese language lesson in the background. Mom Me listens intently to a story about playground politics. She drives carefully and doesn’t complain when turning on JoJo Siwa for the hundredth time. She could never summon enough drama to become the protagonist in any story.
That version, to my child, is the only version of me that matters. And at her young age, that makes sense. She’s not quite ready to see the me beyond her, much less the AI version of me.
But in another life, couldn’t the AI version have been me? If I had made different choices—not gone to graduate school in Chicago, where I met her father; devoted my life to kung fu; been born into a military family predestined for greatness—could I have been a hero, not of my own story, but of all the stories? The AI hero filter is but a small glimpse of another offshoot in the multiverse where I am a different, bolder version of myself. The pull of an alternate self is intoxicating and bewildering. It’s the stuff of movies.
In the film Everything Everywhere All at Once, a struggling, exhausted Evelyn Wang (played by my AI doppelgänger, Michelle Yeoh) learns to navigate the multiverse through verse-jumping technology. Her mission is to save the multiverse by defeating a chaotic, life-destroying being called Jobu Tupaki, who travels fluidly between worlds. To do so, Evelyn must temporarily inhabit the lives of the alternate Evelyns, acquiring their skills in order to reshape her reality. From an opera diva, she learns to reach the highest notes, discombobulating her enemies. From a kung fu fighter, she learns to slice the air with her powerful limbs. From a bizarre yet endearing multiverse where she has hotdogs for fingers, Evelyn learns compassion and vulnerability.
Throughout the film, Evelyn asks several versions of “Why me?” Her guide, an alternate version of her husband Waymond, tells her that he thinks she’s special, that, truly, what makes her so exceptional is her complete ordinariness. It’s not stated explicitly, but the reason Evelyn is able to deftly appropriate so many skills is because she is a blank canvas, a sponge capable of soaking up all the many identities. Until, of course, she isn’t. Until the underlying promise of heroism—the tragic and inevitable martyrdom—catches up with her.
Studying human conversations isn’t a simple challenge. For instance, when humans start to talk to one another in a conversation, they coordinate their speech very tightly—people very rarely talk over one another, and they rarely leave long, unspoken, silent gaps. A conversation is like a dance with no choreography and no music—spontaneous but structured. To support this coordination, the people having the conversation begin to align their breath, their eye gaze, their speech melody and their gestures.
To understand this complexity, studying research participants in a lab looking at computer screens—the traditional setup of psychology experiments—isn’t enough. We need to study how people behave naturally in the real world, using novel measurement techniques that allow us to capture their neural and physiological responses. For instance, Antonia Hamilton, a neuroscientist at University College Londond, has recently used motion capture to identify a pattern of very rapid nods that listeners make to show that they are paying attention when someone is speaking. Hamilton shows that the interaction is improved by these subtle signals, but what’s also fascinating is that although the speakers can actually perceive this information, these body signals are not discernible to the naked eye.
In 2023, we will also finally be able to start capturing neural data while people are moving and talking to each other. This isn’t easy: Brain imaging techniques such as functional magnetic resonance imaging (fMRI) involve inserting participants inside 12-ton brain scanners. A recent study, however, managed that with a cohort of autistic participants. This paper represents a terrific achievement, but, of course, until fMRI techniques become much smaller and more mobile, it is not going to be possible to see how the neural data relates to the pattern of movements and speech in conversations, ideally between both participants in a conversation. On the other hand, a different technique—called functional near infrared dpectroscopy (fNIRS)—can be used while people move around naturally. fNIRS measures the same index of neural activity as fMRI via optodes, which shine light through the scalp and analyze the reflected light. fNIRS has already been deployed while people performed tasks outdoors in central London, proving that this method can be used to gather neural data in parallel with movement and speech data, while people interact naturally.
In 2023 we will also for the first time be able to look at how this would work in large-group conversations, which tend to reach their limit with around five people. This is, of course, a big challenge, as conversations can be so flexible and open-ended, but it’s essential if we want to understand how the participants’ brains coordinate these finely timed conversational dances.
These breakthroughs will represent great strides in the scientific study of human conversation, one of the most fascinating areas of cognitive neuroscience and psychology. Of course, I’m slightly biased: I have studied human speech perception and production for decades, and I think conversations are where our linguistic, social, and emotional brain processes come together. Conversations are universal, and they are the main way that humans use to manage social interactions and connections. They matter hugely to our mental and our physical health. When we can fully crack the science of conversations, we’ll have come a long way to understanding ourselves.
In 2023, radiologists in hospitals around the world will increasingly use medical images—which include x-rays and CT, MRI, and PET scans—that have been first read and evaluated by AI machines. Gastroenterologists will also be relying on machine vision during colonoscopies and endoscopies to pick up polyps that would otherwise be missed. This progress has been made possible by the extensive validation of “machine eyes”—deep neural networks trained with hundreds of thousands of images that can accurately pick up things human experts can’t.
One of the most exciting new capabilities of AI is to instruct untrained and uninitiated people to acquire medical-grade images through a smartphone. Someone without any medical knowledge will be able to pop an ultrasound transducer into a smartphone’s base and, with a little gel on its tip, instantly acquire high-quality images. The AI algorithm instructs the person to move the transducer up or down, clock- or counterclockwise, and it will automatically capture the image when it meets the objective standard. This will extend the ability to perform medical imaging of most parts of the body (except the brain), anywhere, anytime, and by anyone. Concurrently, algorithms are also being developed for automated accurate interpretations. In 2023, we will see more of this in remote parts of the world, perhaps best exemplifying the potential for AI to reduce health inequities.
The same deep-learning democratization is progressively taking hold for patients as well, who can already be notified by their smartwatch’s algorithm that they have an abnormal heart rhythm (such as atrial fibrillation). In 2023, this will extend to preliminary diagnosis of all skin lesions, urinary tract infections, children’s ear infections, and an increasing number of common conditions that are not life-threatening.
These are the early steps towards a virtual health coach to ideally prevent conditions that a person is at increased risk for manifesting, which in 2023 will be used for managing specific conditions such as diabetes, hypertension, or even depression, with the help of chatbots and human coaches in the background when necessary.
In 2023, clinicians will also be aided by AI in their daily tasks—particularly by being liberated from the job of painstakingly typing medical data into the computer. This burden not only contributes to burnout among physicians, but markedly detracts from the patient interactions. Natural language processing and machine learning now enable synthetic notes to be created automatically from the conversation between doctors and patients at the visit or bedside.
We have seen the beginning of use of AI for remote monitoring, which is already preempting the need for hospitalization for patients with Covid-19 by real-time data capture from wearable sensors. That will only increase in 2023. We still need more validation trials to show that algorithms can accurately anticipate early signs of clinical deterioration and intervene, but the implication for avoiding a large proportion of hospital stays looms large.
Nevertheless, there remains a dire need to reduce bias and promote privacy and security in the application of medical AI. Privacy AI computing is starting to take off with the use of federated and swarm learning, as well as with the increasing application of edge computing, which uses algorithms fully operating on the smartphone. In 2023, these strategies will be explored more fully, in a much-needed effort to not only fully investigate the potential for AI in health and medicine but also to address its potential flaws and pitfalls.
In the concrete rubble on Kanokupolu beach, Tonga, leaves have begun forming a cover—green and glossy amid the dull grays of the detritus in the sand. A year after the eruption of Hunga Tonga–Hunga Haʻapai—a volcanic blast bigger than Krakatoa that caused a spike in global warming, reshaped the ocean floor, and wiped out two of the archipelago’s smaller islands—the devastation it wrought is still visible, along with the wreckage of vacation resorts that once stood here, a repair job that is yet to begin.
Last year’s catastrophe, which affected some 84 percent of the Tongan population, was the Pacific nation’s third natural disaster in five years (it was hit by category 5 cyclones Gita and Harold in 2018 and 2020)—a byproduct of global emissions warming the planet, which intensifies storms and droughts, increases wind speeds, and causes sea levels to rise, raising the risk to nearby populations. While coming in at 190th on the global carbon-emission rankings (the US is second), Tonga is now one of many countries being battered by those on distant, richer shores, and being left to pick up the pieces. Aware of this grim fate being meted out to poor nations globally, conversations on how to redress the injustice have begun, largely boiling down to one solution: climate reparations.
A “historic deal” was struck at the Cop27 climate summit in Egypt earlier this year, with the promise to establish a fund that would compensate afflicted countries. Recommendations are due to be made at Cop28 (held in Dubai, ranked 28th for global CO2 emissions) at the end of this year. However, the details remain wooly on how or when they will come into effect. In their absence, it is hard to see the UN’s proposed fund as anything but a hastily applied band-aid designed to assuage the guilty consciences of rich countries without grasping how to truly help those in need, or halt the issues causing these disasters in the first place. As Tonga has found, being repeatedly lashed by the elements requires far more planning and input into prevention than just a hasty clean-up job.
The country needs help, certainly. But having rich nations write a check is not enough. What Tonga (and countries like it) requires is crisis managers who have faced similar disruptions and are skilled at rebuilding communities, and boots on the ground to ensure the money goes where it is truly needed. In the immediate aftermath of last year’s eruption, some nations were quick to send resources, but they rarely matched up with the country’s needs, locals told me when I visited last month. Mounds of food, for instance, when the shops were full of it, were stacked up in a line of ships at the wharf in Nuku’alofa, the capital, delaying other more urgent supplies that then took days to unload. Other gifted items—trucks, clothes—were never even handed out.
Managing these well-intentioned arrivals was nearly impossible with so many more urgent issues to get on top of—like building units for the former residents of Mango and Atata islands, all of whom were evacuated after their homes were destroyed. The first residents were only able to move in just before Christmas. This is a best-case scenario of what climate reparations would look like, in that the new builds solve a direct need, for which on-the ground knowledge and understanding was crucial in both planning and execution. But while these homes are an upgrade on the community halls they were living in for 11 months after the blast, there is no escaping the fact that many now live as 10 family members across two rooms, that they lost their jobs in resorts that were wiped out, and that had sufficient action on climate change been taken sooner, they would not now feel, as one mother told me, as though they had been left with nothing. Their only recourse now is to simply hope another disaster doesn’t strike.
The concern, of course, is that one will—and soon. The Pacific especially is at risk: Kiribati, an idyllic atoll nation between Hawaii and Australia, has in recent years found itself being swallowed by the sea at such a pace that it will likely no longer exist in a few decades. Half of all households have been affected by rising sea levels, with six villages already entirely relocated. The Maldives, Micronesia, and Tuvalu too are predicted to disappear within our lifetimes, with soaring emissions responsible for the coastal erosion, destruction of plantations (and livelihoods), and severe droughts and flooding they and other vulnerable nations routinely face. Larger, wealthier islands like Fiji can’t be immune to the threat either, when 65 percent of its population lives within 5 kilometers of the shore.
Ten years after its discovery, the implications of Crispr genome editing are profound and far-reaching, and we are only getting started. This tool, adapted from a bacterial immune system, allows us to cut and edit the genetic code in any living cell to make highly targeted changes and repairs. A small number of people with genetic diseases have been helped by Crispr therapies, highlighting the potential to impact the lives of those suffering from the approximately 7,000 genetic diseases with known causes. Trials are ongoing in diseases ranging from diabetes to infectious disease.
In 2023, we will begin to benefit from new Crispr-based solutions in other areas. For instance, following on the heels of the initial clinical trial results, the first agricultural applications using Crispr have recently entered the market: A US Food and Drug Administration–approved edit to cattle genes re-creates a slick coat that is occasionally found in nature and allows cows to tolerate increasing temperatures; a Crispr-edited tomato, approved for sale in Japan, has enhanced nutritional qualities. In other crops, Crispr is being used experimentally to increase yield, reduce pesticide and water use, and protect against disease.
The next space for Crispr innovations will be climate change, the defining fight of our times. In 2023, bold new efforts using Crispr to target climate change will begin.
First, new research aims at reducing carbon emissions from agriculture. Agriculture is responsible for about a quarter of all greenhouse gas emissions, and these mainly come from microbes found either in soil, for example in rice paddies, or in the guts of farm animals. This new research is focusing on how to use Crispr to edit these microbes or shift the composition of microbial communities to reduce or even eliminate greenhouse gas emissions.
Second, we are finding ways to improve the inherent ability of plants and microbes to capture carbon and store it in the soil. Plants “breathe in” carbon dioxide during photosynthesis and use it to make energy, but usually the carbon is cycled back into the atmosphere fairly quickly. New research aims to work with plants and soil microbes to not just capture carbon but also store it in soil for long periods of time, replacing some of the soil carbon that has been lost in vast quantities since the advent of modern agriculture.
Third, we are developing new ways to minimize farmer inputs like fertilizers and pesticides that have high carbon costs, as well as other environmental health costs. New Crispr research aims to edit staple crops like rice so they can grow with less fertilizer. Crispr can be used to make plants resistant to common pathogens and pests, reducing the need for high-carbon-emission chemical inputs.
Finally, we need ways to help agriculture deal with the degree of climate change that has already occurred or is inevitable. New research is using Crispr to engineer plants that can produce more food and other materials with less water and that are tolerant to temperature extremes.
A great deal of the attention surrounding Crispr has focused on the medical applications, and for good reason: The results are promising, and the personal stories are uplifting, offering hope to many who have suffered from long-neglected genetic diseases. In 2023, as Crispr moves into agriculture and climate, we will have the opportunity to radically improve human health in a holistic way that can better safeguard our society and enable millions of people around the world to flourish.