The top ten figures of Nature in 2023 are released: ten news figures of the year and one AI

Nature Cover

The pictures and characters in this article are from Nature

ChatGPT is not a person. Yet in many ways, this program has had a profound and widespread impact on science over the past year.

Today, the top scientific journal Nature announced the 2023 list of Nature's 10 scientists - a list that aims to select 10 people who have a place in all major scientific events of the year.

It is worth noting that there is another non-human selected for Nature's 10 this year , that is ChatGPT, the artificial intelligence (AI) assistant that has dominated various news headlines since last year.

Richard Monastersky, editor-in-chief of Nature's features department, said, "Although this tool is not a person and does not fully meet the selection criteria for Nature's top ten people, we have made an exception and included it in the list to recognize the tremendous changes that generative AI has brought to the development and progress of science ."

Nature's 10 explores the important developments in science this year and the ten people who played a major role in these milestones. Together with their colleagues, these people have helped make amazing discoveries and brought attention to critical issues. Among them:

Kalpana Kalahasti played a key role in ensuring the successful landing of Chandrayaan-3 on the Moon, making India the fourth country to achieve such a feat.

Marina Silva helped control rampant deforestation in the Brazilian Amazon and rebuild institutions weakened by the previous government.

Katsuhiko Hayash has created viable eggs using cells from male mice, work that could help save the species from extinction.

Annie Kritcher helps NIF produce nuclear reactions once seen only in hydrogen bombs and stars.

As the United Nations’ Chief Heat Officer, Eleni Myrivili is helping the world prepare for the threats of climate change.

Ilya Sutskever, the pioneer of ChatGPT and other society-changing AI systems, is worried about the future of AI.

James Hamlin helped discover the flaws in the sensational claims of room-temperature superconductivity.

Svetlana Mojsov is gaining recognition for her role in the development of a multi-billion dollar weight loss drug.

Halidou Tinto has helped bring about a second vaccine for a deadly disease .

Thomas Powles leads a transformative clinical trial to treat a severe form of bladder cancer. “The story of these 10 people and one AI tool encapsulates the most important advances in science in 2023,” Monastersky said.

The following is a detailed introduction to the top ten people selected by Nature's 10 and ChatGPT. (The introduction of this article is based on Nature reports and authoritative public information. If there are any omissions, please leave a message to criticize and correct them.)

ChatGPT: A benefit or a burden?

It’s the poster child for generative AI software, and it’s heralding a potential new era of science.

It co-authors scientific papers with humans—sometimes secretly. It drafts outlines for talks, grant proposals, and courses, writes computer code, and serves as a "sounding board" for research ideas. It also makes up references, fabricates facts, and even engages in hate speech. But most of all, it captures people's imaginations : ChatGPT is sometimes obedient, sometimes fascinating, sometimes entertaining, and even terrifying. It plays a variety of roles—some of which users expect, some of which they don't.

Why include a computer program on a list of people who will have a major impact on science in 2023? ChatGPT is not a person. Yet in many ways, the program has had a profound and wide-ranging impact on science over the past year. ChatGPT's sole goal is to credibly continue a conversation in the style of its training data. But in doing so, it and other generative AI programs are changing the way scientists do their work. They're also reigniting debates about the limits of artificial intelligence, the nature of human intelligence, and how best to regulate the interaction between the two. That's why this year's "Nature's 10" includes a new, non-human member.

Some scientists have long recognized the potential of large language models (LLMs). But for many, it was ChatGPT, released as a free conversational agent in November 2022, that quickly became clear how powerful and potentially risky the technology is. The program was created by OpenAI researchers in San Francisco, California, including Ilya Sutskever, who was also selected for "Nature's 10" this year. It is built on a neural network with hundreds of billions of parameters, and its training cost is estimated to be tens of millions of dollars. The training data includes a huge corpus of online books and documents. In addition, they have hired a large number of annotation engineers to edit or evaluate its responses, further constraining the output of the robot. This year, OpenAI upgraded ChatGPT's underlying model and connected it to other programs, enabling the tool to receive and create images and use mathematical and coding software to provide assistance. Other companies have also hurried to launch competing products.

For some researchers, these apps have become invaluable lab assistants—helping summarize or write manuscripts, polishing applications, and writing code. ChatGPT and related software can help people brainstorm, power scientific search engines, and identify research gaps in the literature, says Marinka Zitnik of Harvard Medical School, who works on AI for medical research in Boston. Models trained on scientific data in a similar way could help build AI systems that guide research, perhaps designing new molecules or simulating cellular behavior.

But this technology also has dangers. Automated conversational agents can help cheaters and plagiarists; if not controlled, they may irreversibly pollute the ocean of scientific knowledge. Undeclared AI-generated content has begun to spread on the Internet, and some scientists have admitted to using ChatGPT to generate articles without declaring it.

There’s also the problem of errors and bias, which is inherent to the way generative AI works. Large language models build a model of the world by mapping the interrelationships of language, then spit out possible samples from this distribution without evaluation, without being able to tell what’s true or false. This leads the program to replicate historical biases or inaccuracies in its training data and to make up information, including scientific references that don’t exist.

Emily Bender, a computational linguist at the University of Washington in Seattle, believes there are few appropriate ways to use computer programs or systems that can automatically generate text. ChatGPT has a huge impact on the environment, has problematic biases, and could mislead users into thinking its output comes from a human, she said. On top of that, OpenAI is being sued for stealing data and accused of labor exploitation (hiring freelancers at low wages).

The size and complexity of large language models means they are essentially a “black box,” but when their code and training materials are not publicly available (as is the case with ChatGPT), it can be difficult to understand why they produce certain content. The open-source large model movement is growing, but so far these models are less capable than large proprietary programs like ChatGPT.

Some countries are developing national AI research resources to enable scientists to build and study large-scale generative AI outside of large companies. But it is unclear to what extent regulation will force developers of large models to disclose proprietary information or build in safety features.

No one knows how much more potential ChatGPT-like systems have. Their capabilities may be limited by computing power or new training data. But the generative AI revolution has begun, and there is no turning back .

——By Richard Van Noorden, Richard Webb

Kalpana Kalahasti: Landing on the Moon

The engineer and manager played a key role in ensuring the successful landing of Chandrayaan-3 on the moon, making India the fourth country to achieve the feat.

"We have achieved our goal perfectly," said Kalpana Kalahasti soon after the Indian Space Research Organisation's (ISRO) Chandrayaan-3 mission successfully landed on the moon. "This will be the most memorable and happiest moment for all of us," she added.

As deputy project director of the mission, Kalahasti played a key role in ensuring the success of the mission, a feat that made India the fourth country to successfully land on the lunar surface.

On July 14, Chandrayaan-3 was launched carrying the hopes and anxieties of the entire nation of India. India’s previous attempt, the Chandrayaan-2 mission, failed in 2019 when the lander crashed. In the same year, SpaceIL’s Beresheet lander, and this year Japan’s ispace’s HAKUTO-R mission 1 and Russia’s Luna 25 lander also suffered similar fates.

For Kalahasti and her team, the Chandrayaan-2 lander failure was a critical moment, and they poured all their efforts into preventing the next failure .

One of the biggest challenges the team faced was that the total mass of the spacecraft and the budget had to be consistent with Chandrayaan-2. This meant the team couldn’t significantly redesign the lander or add many redundant systems. So Kalahasti, along with project director Palanivel Veeramuthuvel, reconfigured the orbiter and lander for the Chandrayaan-2 mission. ISRO reduced the mass of the orbiter and gave the lander extra fuel, sturdier legs, and other improvements.

"The experience gained from Chandrayaan-2 was invaluable," Kalahasti told Nature. "Many of its systems worked successfully, allowing us to find the optimal configuration for Chandrayaan-3."

Veeramuthuvel and Kalahasti spent a lot of time on Chandrayaan-3 development, conducting comprehensive tests and simulations, such as evaluating the navigation system's ability to avoid obstacles on lunar-like terrain.

“Our goal was to have a well-documented, well-understood system,” Kalahasti said. “There was no compromise in terms of demonstrating the performance of the system.”

Their efforts paid off. But conducting so many tests and integrating the results while planning the flight was a massive undertaking, requiring coordination among more than a dozen ISRO centers across India. “It was like we were building five or six different satellites together,” Kalahasti said. She drew on past experience in project management and systems engineering, including leadership roles in the development of several of ISRO’s Earth observation satellites.

She was still a long way from leading a lunar mission when she joined ISRO in 2000. She said she was drawn to ISRO because she wanted to work at a core engineering institution that would put her degree in electronics and communications to good use. Her first job at ISRO was as a radar engineer at the Satish Dhawan Space Centre in Sriharikota, where ISRO launches missions.

Kalahasti is heartened by the enthusiasm that the Chandrayaan-3 mission has inspired among the youth of India. “Beyond the technical aspects of the mission, I hope young professionals in India and around the world will be inspired by how the team has meticulously risen from failure.”

The mission's success has also inspired other countries and companies to make future attempts at the moon. "It's great to see India return to this mission so soon after their first attempt," said Jessy Kate Schingler, a space policy researcher and senior advisor at the Open Lunar Foundation. "Moon landings are a very hard thing to do. So Chandrayaan-3, I think, is an appreciated investment that the whole world will benefit from."

Kalahasti is excited about what ISRO might take on next. The agency hopes to send a mission to recover lunar samples as a prelude to its goal of sending humans to the lunar surface by 2040. “Now that the critical lunar landing demonstration is complete, we can move toward other capabilities,” she said.

——By Jatan Mehta

Marina Silva: Protector of the Amazon

Brazil's environment minister has helped control rampant deforestation and rebuild institutions weakened by the previous government.

In a year filled with bad environmental news, with record global warming, scorching heat waves and raging fires, Marina Silva, Brazil’s Minister of Environment and Climate Change, delivered a hopeful message on August 3. She announced that, based on satellite imagery, deforestation alerts in the Amazon rainforest fell 43% from January to July 2023 compared to the same period in 2022. This was in stark contrast to the previous four years, when deforestation alerts had increased significantly.

This shift towards environmental protection in Brazil began on January 1, when Luiz Inácio Lula da Silva assumed the presidency and Marina Silva began her current post. This is her second time leading the Ministry of Environment and Climate Change, a post she previously held during Lula da Silva’s first and second presidential terms, from 2003 to 2008.

During her first tenure, Marina Silva successfully reduced deforestation in the Brazilian Amazon by 83% between 2004 and 2012 by leading the development of the Action Plan for the Prevention and Control of Amazon Deforestation (PPCDAm).

But many of the protections she helped establish were dismantled during Jair Bolsonaro’s presidency from 2019 to 2022. During Bolsonaro’s term, government fines for environmental crimes were cut by 40%, and logging in the Amazon increased by about 60% compared with the four years before.

Marina Silva and her team this year began the “difficult task of rebuilding what was abolished and at the same time creating new outcomes for environmental policy”.

Marina Silva faced many difficulties at an early age. She was born in 1958 in Rio Branco, in the heart of the Amazon region, into a poor family of 11 children (three of whom died young). She and her siblings worked from an early age, extracting latex from rubber trees. She wanted to become a nun and didn't learn to read and write until she was a teenager.

Marina Silva began her environmental activism in the mid-1970s, when she met environmental activist Chico Mendes (who was killed by ranchers in 1988) at a rural leadership course. In 1994, at age 35, she became the youngest elected senator in Brazil.

In her current role, Marina Silva has not always been in line with the current government, according to Pedro Jacobi, an environmental governance researcher at the University of São Paulo in Brazil. Lula da Silva’s government intends to increase oil and gas drilling — including at the mouth of the Amazon. As a result, the environment ministry “has been walking on thin ice,” he said.

But Brazil is making progress in curbing and preventing deforestation, said Natalie Unterstell, director of the Talanoa Institute, a climate policy group in Rio de Janeiro. “Marina Silva’s leadership on this agenda is very important and she’s doing an outstanding job,” she said.

One of her key achievements was the launch on June 5 of an upgraded version of the PPCDam program (which had been shut down by the Bolsonaro government) to protect the Amazon. Marina Silva also restored policing support to the region to enforce environmental regulations. It didn’t take long for these policies to start working. From January to July 2023, fines for environmental crimes issued by the Brazilian Institute of Environment and Renewable Natural Resources (IBAMA) increased by 147% compared to the average for the same period from 2019 to 2022.

According to Brazil’s National Institute for Space Research (INPE), deforestation in the Amazon is expected to be 22% lower between August 2022 and July 2023 than in the previous 12 months. The rate is the lowest since 2018, but still double the lowest deforestation rate in 2012.

Marina Silva said a key reason environmental protection is working is that the agenda is being promoted across all sectors of government. “It fills me with joy to see in practice a concept that I hold very dear – that environmental policy should not be confined to one sector but should run across all sectors,” she said.

However, simply ending deforestation is not enough. "If countries do not reduce carbon dioxide emissions from fossil fuels, forests may also be destroyed in the same way due to climate change. Therefore, we need to change the way we live."

She compares herself to the tough fibers of Amazonian trees, which are used to bind wood to make rafts. “That’s how I see my work, to bring together those who are willing to participate and everything that is needed to form a support surface in the journey of the challenges of our time,” she says.

——By Meghie Rodrigues

Katsuhiko Hayashi: Redesigning Reproduction

He created viable eggs from cells taken from male mice, work that could help save the species from extinction.

When Katsuhiko Hayashi and colleagues announced in March that they had created mouse pups from cells from two male mice, the news stunned some researchers. “I fell off my chair,” says Robert Gilchrist, a reproductive biologist at the University of New South Wales in Sydney. “This is an amazing scientific achievement.”

It’s the culmination of years of hard work by Hayashi, now a developmental biologist at Osaka University in Japan, who and his team previously developed a method to coax mouse stem cells into becoming immature egg cells, then get them to mature, fertilize, and produce live pups (O. Hikabe et al. Nature 539, 299–303; 2016). This year, Hayashi revealed that his lab had successfully created eggs using cells from male mice (K. Murakami et al. Nature 615, 900–906; 2023), a task many researchers had thought was nearly impossible.

Hayashi, however, is more conservative about his achievement: “It wasn’t that hard, actually. ”

Hayashi and colleagues took cells from the tails of male mice that have both X and Y chromosomes and transformed them into stem cells. In the process, about 3% of the cells spontaneously lost their Y chromosomes. The team then isolated these cells without Y chromosomes and treated them with a chemical that causes errors when cells divide.

These errors caused some cells to produce duplicate X chromosomes, effectively turning them into female cells. The team then put these cells through a complex and laborious process to make eggs. Each step is very delicate, and many cells are usually lost, says Mitinori Saitou, a developmental biologist at Kyoto University in Japan who collaborated with Hayashi.

The team then fertilized the eggs and implanted the resulting embryos into female mice. Only seven live pups were produced from 630 embryo transfers.

Over the years, Hayashi has become known for taking on assignments that are both difficult and imaginative. “I admire his work,” says Azim Surani, a developmental biologist and former advisor to Hayashi at the University of Cambridge in the United Kingdom. “His work is very original.”

Collaborators, past and present, have also mentioned Hayashi's patience and work ethic. "He works as fast as 10 people," says Daisuke Kitamura, an immunologist at the Tokyo University of Science, who supervised Hayashi's doctoral research.

Hayashi, for his part, doesn’t get lost in the praise: “It’s really hard for me to get everything done,” he said. “Sometimes I can’t finish my work on time.”

Hayashi had wanted to study germ cells — the cells that naturally give rise to sperm and eggs — since his undergraduate days, intrigued by their importance in reproduction and how they perpetuate life in future generations. “The germ line is the source of life,” he said. “It’s eternal. It’s the only cell line that can survive indefinitely after differentiation.”

Now, his lab is looking to expand this research from mice to another animal: the northern white rhino (Ceratotherium simum cottoni). Today, only two northern white rhinos are known to be alive, both female.

Hayashi's technique may be the way to save the species, but he says it's much harder to grow rhino sperm and eggs in the lab than it is with mice.

The complications are even greater when it comes to dealing with human germ cells. Amander Clark, a developmental and stem-cell biologist at the University of California, Los Angeles, says that the techniques Hayashi has pioneered to make eggs and sperm are likely to take decades before they can be used in humans. In the meantime, Hayashi has largely stayed out of the ethical discussions his work has raised. “From a scientist’s perspective, we have a relatively simple approach: We strive to produce high-quality eggs ,” he says. “ However, it is not our decision whether those eggs should be used. That decision should be made by society. ”

——By Heidi Ledford

Annie Kritcher: Fusion Ignitor

The physicist helped NIF produce nuclear reactions once seen only in hydrogen bombs and stars.

In 2023, physicist Annie Kritcher keeps moving forward with optimism.

A few weeks ago, she helped the U.S. Department of Energy’s National Ignition Facility (NIF) achieve a goal that has long eluded laboratories around the world: fusing atoms together through highly compressed nuclei and producing more energy than the reaction consumes.

However, after reaching this experimental milestone, known as "ignition," the pressure came to achieve more.

The $3.5 billion NIF, based at Lawrence Livermore National Laboratory in California, was designed to enhance nuclear weapons science. Advances could also help advance nuclear fusion as a safe, clean and virtually unlimited energy source. NIF’s successful experiment last year took many by surprise. The “ignition” experiment was a decade behind schedule, and some had begun to worry that the goal was simply out of reach. As lead architect of the major fusion experiment, Kritcher and her team immediately set out to prove that NIF could reliably achieve ignition.

High-risk research doesn’t always go smoothly: The team’s first replication attempt in June was disappointing. “It can be crazy, and I get really stressed out,” Kritcher says.

Fortunately, the next attempt was successful. On July 30, the facility's 192 lasers delivered 2.05 megajoules of energy to a pellet of frozen hydrogen isotopes, deuterium and tritium, contained in a gold cylinder. The resulting implosion released energy as the isotopes fused into helium, generating temperatures six times that of the sun's core. The reactions produced a record 3.88 megajoules of fusion energy.

Other facilities have produced more fusion energy for longer periods of time, particularly in tokamaks, which use powerful magnetic fields to confine fusion reactions. Such technology is being developed at the $22 billion International Thermonuclear Experimental Reactor project, an international collaboration. However, until NIF’s achievement, no laboratory had ever produced a fusion reaction that produced more energy than it consumed.

Kritcher and her team followed up July’s success with two more ignition attempts in October, succeeding four out of six attempts. They are preparing for even higher production capacity next year. In the process, scientists at the facility are opening the door to research and increasing optimism about the future of fusion energy.

NIF is designed to help government scientists ensure the safety and reliability of the U.S. nuclear arsenal without test explosions, but that’s not what initially attracted Kritcher to the lab. She began studying fusion energy as a summer intern at Livermore in 2004, and before starting graduate school, she quickly set her sights on NIF because it’s one of the few places on Earth studying fusion reactions.

She joined NIF in 2012 and became chief designer in 2016. Since then, she has led a team to analyze experimental data and use computer models to design experiments aimed at achieving and improving nuclear fusion yields by adjusting parameters such as the size and configuration of the target and the energy and timing of various laser beams. Once her team completes the design, the laboratory's experimental team will take over to fire the lasers and collect data.

“Kritcher was an excellent student who threw herself into her work,” said Roger Falcone, a physicist at the University of California, Berkeley, who has worked with Kritcher since she was a graduate student and through her early work at NIF. During that time, she demonstrated her prowess in designing laser experiments to test how materials behave under extreme temperatures and pressures.

Those skills put Kritcher at the center of the fusion program in 2016. Energy output from fusion experiments had stagnated, and NIF chief scientist Omar Hurricane wanted a new approach. Kritcher came up with some ideas, and “she threw herself into them,” Hurricane says, and that’s how she became one of NIF’s lead designers.

Kritcher and her team spent the next few years doing a lot of number crunching and design tweaking on what became the main NIF experiment. In addition to making various changes to the target, they also used various improvements to increase the overall energy of the laser. As a result, they achieved fusion more and more often .

Now that Kritcher has achieved the official ignition goal, she is working on a series of new experiments aimed at increasing the throughput again by pumping more laser energy into thicker target capsules. This represents another step toward NIF’s goal of achieving throughput of tens of megajoules or more.

In the long term, she believes the facility, after some upgrades, will be able to achieve its goals and increase output by an order of magnitude, which would allow scientists to hopefully begin working on prototype laser fusion energy reactors. “It’s not a question of ‘if,’ it’s a question of ‘when,’ ” she said, in which she hopes lasers will play a role.

“I think it’s a great possibility,” she said, “and I’d love to be a part of it.”

——By Jeff Tollefson

Eleni Myrivili: Climate Change Supervisor

As the UN's chief heat officer, the former politician is helping the world prepare for the threat of climate change.

As temperatures soared and wildfires raged across Greece in July, Eleni Myrivili watched in horror from her home in Athens. “This kind of destruction is irreplaceable,” she said. “This is clearly a disaster for which we have to take responsibility.”

Myrivili, a former deputy mayor of Athens who now works around the world to mitigate the catastrophic effects of a warming climate, is the United Nations’ first chief heat officer, charged with keeping people cool as the planet “boils.”

July broke unofficial records for the hottest daily and monthly global average temperatures , and by the end of December, 2023 will almost certainly be the hottest year on record. Extreme heat is taking a deadly toll. A paper published in July estimated that the 2022 European heatwave killed more than 61,000 people between late May and early September, with the highest heat-related mortality in Italy, Greece, and Spain (J. Ballester et al. Nature Med. 29, 1857–1866; 2023).

Myrivili has been interested in heat for years, beginning with her career as a cultural anthropologist. She earned a PhD studying migration, violence and border regions, specializing in the region where Albania, North Macedonia and Greece meet. She then began teaching at the University of the Aegean in Greece. In the summer of 2007, fires burned across parts of Greece, destroying much of Parnis National Park near Athens. “All of us who work on climate change have that moment when the reality of climate change hits you,” she says. “For me, it was 2007.”

Angry at the lack of information and advice about the fires, Myrivili decided to go into politics. After a stint working for a Greek green party, she read Benjamin R. Barber’s 2013 book If Mayors Ruled the World and decided to turn to city government.

She has held various positions in the Athens government, working to integrate climate resilience and planning into various city departments. When she became head of the parks department, she changed her title to head of the city’s nature, resilience and climate change adaptation to emphasize what she sees as her focus. In 2021, she was named Athens’ Chief Heat Officer.

To raise awareness of deadly weather patterns, Myrivili has launched a campaign to name heat waves in Greece. She has also worked to secure funding for climate initiatives; in 2018, her group received a €5 million loan (then $5.9 million) from the European Investment Bank for climate adaptation projects, including planting green spaces in Athens.

Elissavet Bargianni, who succeeded Myrivili as Athens' chief heat officer earlier this year, said Myrivili's efforts and connections have enabled her to gain effective support within the city government. "A lot of people trust her," Bargianni said. "She can make the seemingly impossible possible."

Myrivili’s work is now focused on raising awareness of extreme heat around the world and securing funding for projects through UN-HABITAT, a program dedicated to sustainable urban development. This includes the Global Cooling Commitment, presented at the COP28 climate conference in Dubai this month. The pledge aims to support the development and deployment of cooling technologies that do not increase greenhouse gas emissions.

Don’t mention air conditioning to Myrivili, who hates its impact on the climate. But during a heat wave in 2021, she finally bought a small unit for her bedroom. “I still hate it,” she says.

However, as temperatures climb, air conditioning use will only increase. That’s one of the reasons Myrivili is attending the Dubai climate talks, to stop the world from burning up and prepare for the future.

——By Alexandra Witze

Ilya Sutskever: AI Visionary

He is a pioneer of ChatGPT and other society-changing AI systems.

As a teenager, looking for a job, Ilya Sutskever knocked on the door of modern artificial intelligence (AI) godfather Geoffrey Hinton at the University of Toronto in Canada. "He said he would rather work for me, working on AI, than fry French fries on his vacation for money, " Hinton said.

Hinton gave his ambitious student some papers to read, and Sutskever came back wondering why the authors hadn’t worked out what seemed to him to be an obvious solution . “His intuition was right,” says Hinton, who considers Sutskever a visionary pioneer in deep learning and large language models (LLMs), the foundation for conversational AI bots like ChatGPT, which took the world by storm this year. “What makes him special is not just his intelligence,” Hinton says, “but the urgency with which he approaches things.”

Sutskever later became OpenAI's chief scientist and played a central role in developing ChatGPT. However, he also worries about the future of AI. In July, he shifted his focus to co-leading OpenAI's four-year "superalignment" project. OpenAI said the project will use 20% of its computing power to study how to "guide and control AI systems that are much smarter than us."

The tension between safety and the commercial motivation to move fast may have led to a confusing saga in November, in which Sutskever played a key role in the firing and rehiring of OpenAI’s chief executive, Sam Altman. After the turmoil, Sutskever declined to be interviewed by Nature.

To some, Sutskever’s foresight is admirable. “He has a strong sense of morality,” Hinton said. “He’s really concerned about the safety of AI.” But some say that focusing on how to control AI systems that haven’t arrived yet distracts from the very real dangers of the technology. It “postpones intervention until a distant future,” said Sarah Myers West, managing director of the AI ​​Now Institute, a policy research group in New York City. Instead, she said, we need to “address near-term harms,” such as AI systems that reinforce biases in training data or that could leak private information.

The lack of transparency in AI systems is also a concern. OpenAI and some other companies keep their code and training data secret. Sutskever said that in the long run, closed systems will be the responsible approach to prevent others from creating powerful AI. "There will be a point in time when AI capabilities will become so powerful that it would be irresponsible to open source the model," he said in April.

Born in 1986, Sutskever began taking university-level coding courses as a teenager. After his family moved to Canada, he began working with Hinton on deep learning in 2003. In 2012, Sutskever and another Hinton student (Alex Krizhevsky) built the neural network AlexNet and won a landmark image recognition competition by a stunning margin. Sutskever later went to Google, where he helped develop AlphaGo.

In 2015, Sutskever was invited to dinner with Altman, Musk and others. That same year, they co-founded the nonprofit OpenAI, which aims to "benefit humanity." Sutskever saw this as an opportunity to seriously pursue general artificial intelligence (AGI). "I would say that researchers are somewhat trained to start small... But at OpenAI, we dare to think big ," he said earlier this year.

Wojciech Zaremba, another OpenAI co-founder, credits Sutskever with pushing the company to invest more effort in its generative pre-trained Transformer (GPT) system after the launch of GPT-1 in 2018. Unlike many others at the time, Sutskever was convinced that these systems could be made smarter simply by increasing computing power. “He understood this before almost anyone else,” Hinton said.

To attract more money needed for computing power, the team changed OpenAI from a non-profit model to a "capped profit" model in 2019, attracting tech giant Microsoft to invest billions of dollars in cash and computing resources into it. This pays off: the model has been improved, and their ChatGPT, released in November 2022, caused a sensation around the world.

Just after this success, Sutskever and other OpenAI board members fired Altman on November 17, causing internal chaos. Many employees threatened to join Microsoft with Altman. Later, Sutskever expressed regret for his actions. Five days later, when Altman rejoined the company, Sutskever was removed from the board.

Sutskever's statement about AI has been very bold throughout. In 2022, he claimed that AI may have "had some perception" that has aroused people's amazement, fear and ridicule. He publicly stated that AGI even "super intelligence" that surpasses the sum of human intelligence may be developed in years or decades. "To this day, I am still surprised by his optimism," said AI researcher Anda Ng.

But Andy Ng also said that Sutskever "has an admirable trait, that is, whether others agree with him or not, he can choose a direction and pursue it relentlessly."

——By Nicola Jones

James Hamlin: Superconducting Investigator

The physicist helped to discover the flaws of the sensational claim of room temperature superconductivity.

James Hamlin recalls his first time being deceived by an experiment. As a graduate student, Hamlin discovered a superconducting phenomenon in an unexpected material, the phenomenon of electrons flowing without resistance. He excitedly shared the news with his supervisor, but his supervisor seemed calm. "He asked me a lot of questions and suggested that I take more measurements," Hamlin said. After further inspection, the superconducting signal disappeared. He learned a direct lesson from it: "Don't assume you found something."

The lesson was reflected on this year’s international stage, when physicist Ranga Dias of the University of Rochester, New York reported in Nature that he achieved his long-awaited room temperature superconducting goal under moderate pressure.

Amid a wave of hype and criticism, Hamlin, who conducted a high-pressure experiment at the University of Florida at Gainesville, and Brad Ramshaw, a superconducting researcher at Cornell University in Ithaca, New York, expressed their concerns about the study to Nature. The paper was withdrawn in November, causing a sensation: It was the third time Dias has been withdrawn in more than a year.

Hamlin has repeatedly exposed problems at Dias' work. In 2020, Dias published a paper in Nature claiming to have discovered the first room temperature superconductor, although it needs to be achieved at higher pressures. All known superconductors currently have to work at very low or high pressures. A superconductor that can operate at room temperature and pressure may lead to imaginative applications, such as magnets and efficient computer chips that do not require expensive cooling equipment—these tempting possibilities have led to the hype around room temperature superconductors.

After Dias's 2020 paper was published, physicist Jorge Hirsch of the University of California, San Diego believes that one measurement in the study looks suspicious—and similar to that in a 2009 paper Hamlin co-authored. Driven by Hirsch, Hamlin investigated his work and found that another co-author, Matthew Debessai, manipulated that data. (Debessai, who no longer works in the research, did not respond to a request for comment.) The paper was withdrawn in 2021, but Hamlin suspects Dias's 2020 study was also problematic.

Dias and co-author Ashkan Salamat, a physicist at the University of Nevada, Las Vegas, took more than a year to release the data Hirsch wanted. Analysis by Hirsch, Hamlin and others found evidence of manipulation. In September 2022, Nature withdrew the work; the withdrawal statement did not mention misconduct, and Dias denied allegations of misconduct.

Hamlin also found that Dias's paper copied his own and others' work—Dias also reused some paper data in a later paper by Physical Review Letters. The paper was withdrawn in August this year. (Although Dias disagreed with the withdrawal, he admitted that he did not provide "clear attribution" for some of the paper.)

In March, Hamlin presented his evidence at a virtual workshop, and some observers expressed "shocking" at the work he did, said Brian Skinner, a physicist at Ohio State University and conference organizer. Hamlin at one point created a tool to extract data directly from Dias' charts because he had no access to the original resistive data. "It was very heroic," Skinner said.

This controversial background is why many researchers were surprised by Nature’s second paper published by Dias in March, which also claimed to achieve room temperature superconductivity, despite being in different materials.

This time, a lot of raw data was public and problems quickly emerged online. Ramshaw and Hamlin focused on some core issues, including whether Dias actually measured the resistor to zero.

In subsequent round-trip discussions with Ramshaw, Hamlin and Nature editors, Dias and his co-author Salamat did not explain how they got the measurement. "We couldn't get a clear answer to this very simple question," Hamlin said. The Nature news team contacted Dias and Salamat for comment, but received no responses.

Then, in September, eight of the 11 authors, including Salamat, asked to withdraw the paper, confirming the concerns raised by Hamlin and Ramshaw. Nature withdrew the paper on November 7, due to concerns about data integrity.

Hamlin and Ramshaw said data availability makes the latest retraction easier: this time it took only half a year, not two years.

Shanti Deemyad, who worked as a high-pressure experimenter at the University of Utah Salt Lake City, coached Hamlin in the lab as an undergraduate. She was not surprised by his dedication. “He was very ambitious and very excited,” she said, “he wanted to know all the details.” Even though she came to the lab at 6:30 a.m., Hamlin was there, keeping her passion for learning.

Hamlin is not a full-time investigator, and he is eager to spend more time doing his superconducting research. “It’s still the most exciting topic of physics I think,” he said. “Human nonsense is much more boring than discovering the mysteries of nature.”

——By Dan Garisto

Svetlana Mojsov: Nameless Drug Developer

How a biochemist ended up gaining recognition for being involved in the development of multi-billion dollar weight loss drugs.

A new weight-loss drug category has quickly entered clinics, making drugs like Ozempic and Wegovy a household name. They have brought billions of dollars in profits to the pharmaceutical industry and earned scientific acclaim for researchers who discovered the hormone behind these drugs, a appetite suppressant called glucagon-like peptide-1 (GLP-1).

However, there was an early pioneer who did not receive the recognition it deserved : Svetlana Mojsov. She is now a biochemist at Rockefeller University in New York City, who played a key role in determining and describing the active morphology of GLP-1. However, her efforts in the history of hormone discovery have been ignored in many accounts and have not shared the scientific awards given for their achievements.

This year, Mojsov challenged those deep-rooted claims — and began to gain wider recognition for her contribution to the field. “What I did is just to correct the scientific record,” she said.

The scientist, now in her seventies, was a member of the endocrinology department of Massachusetts General Hospital (MGH) in Boston in the 1980s. She also led an institution that makes synthetic proteins for use inside and outside the department. During this period, she conducted a series of landmark studies and provided others with the tools they needed to conduct their research.

Her work with GLP-1 began with her prediction that a specific version of hormone should be present in the intestinal tissue of a mammal, which she later verified by experiments (S. Mojsov et al. J. Biol. Chem. 261, 11880–11889; 1986). She later demonstrated that this biologically active morphology of GLP-1 triggers insulin release in rat pancreas (S. Mojsov et al. J. Clin. Invest. 79, 616–619; 1987).

The peptides and antibodies created by Mojsov were also crucial to several other GLP-1 experiments conducted in cell lines at the time and allowed clinicians to demonstrate that GLP-1 could lower blood sugar in early human trials.

These studies laid the foundation for drugs such as Ozempic and Wegovy, which contain a GLP-1 analog called semaglutide. It only made minor changes to the peptide outlined in Mojsov’s original paper; these changes improved stability and ensured longer-lasting results. Global sales of semagglutide now exceed $1 billion per month, and this class of drugs is expected to be one of the best sold drugs ever.

However, the role of Mojsov in the discovery process has been long ignored.

She had to go through a long legal battle to add her name to the base patent as a co-inventor, a move that led Mojsov to receive a royalty of one or two years for the sale of the first generation of GLP-1 drugs. But as her patent expired for a long time, she gained no financial gain in the huge profits of semagglutide.

Mojsov began to feel dissatisfied with the lack of recognition. History is being “manipulated”, she said. For example, she believes that comments made to match the awards over-exaggerate the contributions of some of the winners and ignore her efforts .

"That was the moment when I decided to fight back," Mojsov said.

At her urging, journals such as Cell and Nature have revised the narrative of the history of GLP-1 discovery to better reflect Mojsov's involvement in MGH - one of two places that independently focus on and describe the active form of the hormone, the other being the University of Copenhagen. In September 2023, Science magazine and news outlet STAT published two detailed profiles, finally telling her story about 40 years after she began studying GLP-1.

She has since received emails showing support – from peer scientists, especially women who believe her work is equally marginalized, and those frustrated by the hierarchy of biomedical research.

Focusing on Mojsov's attention has also begun to change the minds of some people in early GLP-1 studies.

“She makes sense,” said Joel Habener, a molecular endocrinologist at MGH. Habener worked with Mojsov to appear as a senior author on all of her groundbreaking papers, but he was the sole patent holder until Mojsov corrected the patent. “She definitely deserves recognition,” he said.

In past reports on the discovery, Mojsov was misdescribed as a scientist in Habener’s research group, rather than an independent researcher, whose efforts helped drive the work of MGH. “Her contribution is crucial,” said Richard Goodman, a molecular neurobiologist at Oregon University of Health and Science, Portland. Goodman, a former postdoctoral fellow at Habener, helped decode the genes of the GLP-1 precursor. “Will it move forward without Mojsov? No.”

Although the awards and the reputation they bring may follow, this is not Mojsov’s priority, and she will continue to study GLP-1 and related proteins in the lab.

“I’m just glad my job was recognized,” she said, “everything else is secondary.”

——By Elie Dolgin

Halidou Tinto: Malaria Fighter

Thanks to the researcher's rigorous testing, a second vaccine for the deadly disease is about to come.

In October 2020, when Halido Tinto's six-year-old daughter contracted malaria, his work- and life clashed. Tinto, who has been director of clinical trials for more than a decade as clinical trial of malaria drugs and vaccines, knows the severity of the disease. His daughter was hospitalized for four days with fever, headaches and vomiting, and although she finally recovered, he said "the situation is really serious."

Just the same month, the vaccine R21 he has been testing received recommended use by the World Health Organization (WHO). It is the second approved malaria vaccine, which many believe can prevent millions of deaths in Africa, with more than 200 million cases and 500,000 deaths each year, mainly children under the age of 5 .

The agency led by Tinto, the Clinical Research Unit of Burkina Fasonanoro (CRUN), is a key location for testing R21, its predecessor RTS, S, and several other drugs. Many scientists believe that Tinto's diligence is the key to the agency's success.

Tinto received his PhD from the University of Antwerp in Belgium to study the resistance of malaria to various drugs. His then-tutor Umberto D'Alessandro, a clinical epidemiologist at the London School of Hygiene and Tropical Medicine, commented that Tinto was both rigorous and dedicated as a scientist. “He really wanted to promote science and research in Africa,” D'Alessandro said.

Tinto had the opportunity to do postdoctoral research at a U.S. university, but he chose to return to Burkina Faso in 2006. There, he helped establish CRUN with local scientists and clinicians.

In 2007, pharmaceutical company GSK and its partners prepared for late-stage clinical trials of the RTS,S vaccine—a vaccine developed over the years. To be part of the trial seemed an out of reach for the new Tinto clinic with only 10 employees. “They were surprised by our application,” he said, “because there was no electricity, no cars, nothing.” Still, Tinto convinced the coordinator.

He met the king in the village and persuaded the Burkina Faso government to connect Nanoro to the grid. CRUN provided data to help RTS,S get approved in Africa.

This vaccine is associated with a significant drop in child mortality. But GSK can only produce a few million doses per year. Even if Burkina Faso gets 1 million of these doses, Tinto said that it can only vaccinate 250,000 children a year. "We still have millions of children," he said. That's why people are excited about R21: Currently, the Serum Institute in Pune, India can produce 100 million doses of R21 per year. R21 will be more affordable compared to RTS,S, and some researchers expect it to be more effective.

Tinto has conducted an influential early study of the vaccine since 2019 (MS Datoo et al. Lancet 397, 1809 1818; 2021). “The trials he led really made people in the field think that this vaccine would be different,” said Adrian Hill, a vaccineologist at the University of Oxford in the UK, who oversees the development of R21.

WHO said R21 will be rolled out across Africa by mid-2024. Meanwhile, Tinto is conducting more than 30 clinical trials, including two other malaria vaccines and more studies on R21.

CRUN has surpassed Nanoro and currently has more than 400 employees and partners, including dozens of graduate students from all over Africa. Tinto, who is still working with D'Alessandro, said it is a great example of how to research how to promote Africa. But what motivates Tinto is the chance to save lives. “Nothing is more satisfying than that, because to me, life is the most important.”

——By Brendan Maher

Thomas Powles: Cancer Explorer

The doctor and cancer researcher led a transformative clinical trial for severe bladder cancer.

When Thomas Powles browsing data on a clinical trial of advanced bladder cancer, he could hardly believe the results. "I thought there would always be some problems," said Powles, a cancer researcher who works at St. Bartholomew's Hospital in London. But he didn't.

Compared to traditional chemotherapy, the combination of the two new drugs seems to extend the average survival time of patients from about 16 months to 2.5 years. When Powles released the data at the European Society of Medical Oncology Conference in Madrid in October, the audience burst into applause. "That was really moving," Powles said.

“This trial is undoubtedly our biggest breakthrough in the field of treating advanced bladder cancer in the past 40 years,” said Eila Skinner, a bladder cancer researcher at Stanford University in California. The drug combination is the first drug to surpass the standard treatment effect since the 1980s.

These exciting results are added to an increasing number of studies that highlight the prospects of a newly developed class of treatments—anti-drug conjugates (ADCs). ADCs are made from a specific antibody targeting tumor cells and a toxic chemotherapeutic drug. Studies on other cancer types also show that ADCs can slow tumor growth and prolong the survival of the subjects, Meric-Bernstam said. “I think ADCs will be the next pillar of cancer care. It’s clear that we will be able to change the outcomes of treatment across multiple tumor types. ”

This bladder cancer trial tested an ADC called enfortumab vedotin, which targets a protein called nectin-4, which is very abundant in bladder cancer cells. It is used with pembrolizumab, another immunotherapy drug, which can relieve the inhibition of immune cells against cancer.

Earlier this year, the U.S. Food and Drug Administration approved the therapy for those who are not suitable for receiving some kind of chemotherapy called cisplatin, which includes about half of patients with advanced bladder cancer.

The drug may get approval for wider use early next year, said Guru Sonpavde, a bladder cancer researcher at the Advent Health Cancer Institute in Orlando, Florida.

Powles has almost become a cardiologist after completing her medical training. But he realized there are more opportunities for discovery in the field of bladder cancer. “For 20 years, people have been receiving the same treatment. The life expectancy of patients is 12 months, and unfortunately almost everyone dies of the disease,” he said. “On this map, only a very small part is depicted.”

He has led more than 20 randomized clinical trials, several focusing on applying advanced immunotherapies to kidney and bladder cancer. Many trials have failed to show improvements, but Powles persisted and believes there are ways to improve survival. The biggest challenge is to convince research funders to invest. “Sometimes we have storms on this journey, but even the failed trials, we are discovering something new,” Powles said.

Sonpavde, who has known Powles for more than a decade, said Powles could come up with great ideas and put them into practice. “A lot of people have ideas,” he said, but those ideas “have not been implemented.”

Powles loves traveling and can draw in his spare time. He has been creating a series of aerial urban landscapes, which reflects his love for portraying complex systems. "I can get lost in it. If I don't have a good time, I'll adjust one of them," he said.

But besides the desire for discovery, Powles said, driving him was also the patients he treated. “The patients who participated in these trials really made extraordinary sacrifices,” he added.

——By Carissa Wong

Original link:

https://www.nature.com/articles/d41586-023-03919-1