Nature Climate Change, Published online: 02 May 2025; doi:10.1038/s41558-025-02331-z

Reply to: Individualized cost–benefit analysis does not fit for demand-side mitigation

The right to repair just keeps on winning. Last week, thanks in part to messages from EFF supporters, the Washington legislature passed a strong consumer electronics right-to-repair legislation through both the House and Senate. The bill affirms our right to repair by banning restrictions that keep people and local businesses from accessing the parts, manuals, and tools they need for cheaper, easier repairs. It joined another strong right-to-repair bill for wheelchairs, ensuring folks can access the parts and manuals they need to fix their mobility devices. Both measures now head to Gov. Bob Ferguson. If you’re in Washington State, please urge the governor to sign these important bills.

TAKE ACTION

Washington State has come close to passing strong right-to-repair legislation before, only to falter at the last moments. This year, thanks to the work of our friends at the U.S. Public Interest Research Group (USPIRG) and their affiliate Washington PIRG, a coalition of groups got the bill through the legislature by emphasizing that the right to repair is good for people, good for small business, and good for the environment. Given the cost of new electronic devices is likely to increase, it’s also a pocketbook issue that more lawmakers should get behind.  

This spring marked the first time that all 50 states have considered right-to-repair legislation. Seven states—California, Colorado, Massachusetts, Minnesota, Maine, New York, and Oregon—have right-to-repair laws to date. If you’re in Washington, urge Gov. Ferguson to sign both bills and make your state the eighth to join this elite club. Let’s keep this momentum going!

TAKE ACTION

Simple common sense tells us that a corporation’s decision to operate in every state shouldn’t mean it can’t be sued in most of them. Sadly, U.S. law doesn’t always follow common sense. That’s why we were so pleased with a recent holding from the Ninth Circuit Court of Appeals. Setting a crucial precedent, the court held that consumers can sue national or multinational companies in the consumers’ home courts if those companies violate state data privacy laws.

The case, Briskin v. Shopify, stems from a California resident’s allegations that Shopify, a company that offers back-end support to e-commerce companies around the U.S. and the globe, installed tracking software on his devices without his knowledge or consent, and used it to secretly collect data about him. Shopify also allegedly tracked users’ browsing activities across multiple sites and compiled that information into comprehensive user profiles, complete with financial “risk scores” that companies could use to block users’ future purchases. The Ninth Circuit initially dismissed the lawsuit for lack of personal jurisdiction, ruling that Shopify did not have a close enough connection to California to be fairly sued there. Collecting data on Californians along with millions of other users was not enough; to be sued in California, Shopify had to do something to target Californians in particular.  

Represented by nonprofit Public Citizen, Briskin asked the court to rehear the case en banc (meaning, review by the full court rather than just a three-judge panel). The court agreed and invited further briefing. After that review, the court vacated the earlier holding, agreeing with the plaintiff (and EFF’s argument in a supporting amicus brief) that Shopify’s extensive collection of information from users in other states should not prevent California plaintiffs from having their day in court in their home state.   

The key issue was whether Shopify’s actions were “expressly aimed” at California. Shopify argued that it was “mere happenstance” that its conduct affected a consumer in California, arising from the consumer’s own choices. The Ninth Circuit rejected that theory, noting:

Pre-internet, there would be no doubt that the California courts would have specific personal jurisdiction over a third party who physically entered a Californian’s home by deceptive means to take personal information from the Californian’s files for its own commercial gain. Here, though Shopify’s entry into the state of California is by electronic means, its surreptitious interception of Briskin’s personal identifying information certainly is a relevant contact with the forum state.

The court further noted that the harm in California was not “mere happenstance” because, among other things, Shopify allegedly knew plaintiff's location either prior to or shortly after installing its initial tracking software on his device as well as those of other Californians.

Importantly, the court overruled earlier cases that had suggested that “express aiming” required the plaintiff to show that a company “targeted” California in particular. As the court recognized, such a requirement would have the

perverse effect of allowing a corporation to direct its activities toward all 50 states yet to escape specific personal jurisdiction in each of those states for claims arising from or relating to the relevant contacts in the forum state that injure that state’s residents.

Instead, the question is whether Shopify’s own conduct connected it to California in a meaningful way. The answer was a resounding yes, for multiple reasons:

Shopify knows about its California consumer base, conducts its regular business in California, contacts California residents, interacts with them as an intermediary for its merchants, installs its software onto their devices in California, and continues to track their activities.

In other words, a company can’t deliberately collect a bunch of information about a person in a given state, including where they are located, use that information for its own commercial purposes, and then claim it has little or no relationship with that state.

As states around the country seek to fill the gaps left by Congress’ failure to pass comprehensive data privacy legislation, this ruling helps ensure that those state laws will have real teeth. In an era of ever-increasing corporate surveillance, that’s a crucial win.

Two essays were just published on DOGE’s data collection and aggregation, and how it ends with a modern surveillance state.

It’s good to see this finally being talked about.

The composition of bacterial populations living on our faces plays a significant role in the development of acne and other skin conditions such as eczema. Two species of bacteria predominate in most people, but how they interact with each other, and how those interactions may contribute to disease, has been difficult to study.

MIT researchers have now revealed the dynamics of those interactions in more detail than previously possible, shedding light on when and how new bacterial strains emerge on the skin of the face. Their findings could help guide the development of new treatments for acne and other conditions, and may also help to optimize the timing of such treatments.

The researchers found that many new strains of Cutibacterium acnes, a species believed to contribute to the development of acne, are acquired during the early teenage years. But after that, the makeup of these populations becomes very stable and doesn’t change much even when exposed to new strains.

That suggests that this transitional stage could be the best window for introducing probiotic strains of C. acnes, says Tami Lieberman, an associate professor of civil and environmental engineering, a member of MIT’s Institute for Medical Engineering and Science, and the senior author of the study.

“We found that there are some surprising dynamics, and these dynamics provide insights for how to design probiotic therapy,” Lieberman says. “If we had a strain that we knew could prevent acne, these results would suggest we should make sure we apply them early during the transition to adulthood, to really get them to engraft.”

Jacob Baker PhD ’24, who is now the chief scientific officer at Taxa Technologies, is the lead author of the paper, which appears today in Cell Host and Microbe. Other authors include MIT graduate student Evan Qu, MIT postdoc Christopher Mancuso, Harvard University graduate student A. Delphine Tripp, and former MIT postdoc Arolyn Conwill PhD ’18.

Microbial dynamics

Although C. acnes has been implicated in the development of acne, it is still unclear exactly why acne develops in some people but not others — it may be that some strains are more likely to cause skin inflammation, or there may be differences in how the host immune system responds to the bacteria, Lieberman says. There are probiotic strains of C. acnes now available, which are thought to help prevent acne, but the benefits of these strains have not been proven.

Along with C. acnes, the other predominant bacterium found on the face is Staphylococcus epidermidis. Together, these two strains make up about 80 percent of the strains in the adult facial skin microbiome. Both of these species exist in different strains, or lineages, that vary by a small number of genetic mutations. However, until now, researchers had not been able to accurately measure this diversity or track how it changes over time.

Learning more about those dynamics could help researchers answer key questions that could help them develop new probiotic treatments for acne: How easy is it for new lineages to establish themselves on the skin, and what is the best time to introduce them?

To study these population shifts, the researchers had to measure how individual cells evolve over time. To do that, they began by obtaining microbiome samples from 30 children at a Boston-area school and from 27 of their parents. Studying members of the same family enabled the researchers to analyze the likelihood of different lineages being transferred between people in close contact.

For about half of the individuals, the researchers were able to take samples at multiple time points, and for the rest, only once. For each sample, they isolated individual cells and grew them into colonies, then sequenced their genomes.

This allowed the researchers to learn how many lineages were found on each person, how they changed over time, and how different cells from the same lineage were. From that information, the researchers could infer what had happened to those lineages in the recent past and how long they had been present on the individual.

Overall, the researchers identified a total of 89 C. acnes lineages and 78 S. epidermidis lineages, with up to 11 of each found in each person’s microbiome. Previous work had suggested that in each person’s facial skin microbiome, lineages of these two skin bacteria remain stable over long periods of time, but the MIT team found that these populations are actually more dynamic than previously thought.

“We wanted to know if these communities were truly stable, and if there could be times where they weren’t stable. In particular, if the transition to an adult skin like microbiome would have a higher rate of acquisition of new lineages,” Lieberman says.

During the early teens, an increase in hormone production results in increased oil on the skin, which is a good food source for bacteria. It has previously been shown that during this time, the density of bacteria on the skin of the face increases by about 10,000-fold. In this study, the researchers found that while the composition of C. acnes populations tended to remain very stable over time, the early teenage years present an opportunity for many more lineages of C. acnes to appear.

“For C. acnes, what we were able to show was that people do get strains throughout life, but very rarely,” Lieberman says. “We see the highest rate of influx when teenagers are transitioning to a more adult-like skin microbiome.”

The findings suggest that for topical probiotic treatments for acne, the best time to apply them is during the early teenage years, when there could be more opportunity for probiotic strains to become established.

Population turnover

Later in adulthood, there is a little bit of sharing of C. acnes strains between parents living in the same household, but the rate of turnover in any individual person’s microbiome is still very low, Lieberman says.

The researchers found that S. epidermidis has a much higher turnover rate than C. acnes — each S. epidermidis strain lives on the face for an average of less than two years. However, there was not very much overlap in the S. epidermidis lineages shared by members of the same household, suggesting that transfer of strains between people is not causing the high turnover rate.

“That suggests that something is preventing homogenization between people,” Lieberman says. “It could be host genetics or host behavior, or people using different topicals or different moisturizers, or it could be active restriction of new migrants from the bacteria that are already there at that moment.”

Now that they’ve shown that new C. acnes strains can be acquired during the early teenage years, the researchers hope to study whether the timing of this acquisition affects how the immune system responds to them. They also hope to learn more about how people maintain such different microbiome populations even when exposed to new lineages through close contact with family members.

“We want to understand why we each have unique strain communities despite the fact that there is this constant accessibility and high turnover, specifically for S. epidermidis,” Lieberman says. “What’s driving this constant turnover in S. epidermidis, and what are the implications of these new colonizations for acne during adolescence?”

The research was funded by the MIT Center for Microbiome Informatics and Therapeutics, a Smith Family Foundation Award for Excellence in Biomedical Research, and the National Institutes of Health.

He is the first president in at least three decades to deny governors’ requests for funding that’s meant to protect people and property.

State legislators voted down several bills this week that would have tightened regulations of CCS projects.

Scientists previously have suggested that the hypothetical climate fix would require specialized high-altitude vehicles.

André Aranha Corrêa do Lago will oversee a challenging climate summit in the Amazon. Here’s what he has to say about President Donald Trump, China and future COPs.

New research indicates that rising temperatures threaten companies' assets in 48 countries.

RBC, Canada’s biggest bank, cited changes to the country’s Competition Act that restrict the kinds of environmental claims firms can make.

President Donald Trump's current 10 percent tariffs cover most coffee-producing countries, including Brazil, Ethiopia and Colombia, and are expected to drive up costs for Americans.

The governor has long said the 10 million visitors who come to Hawaii each year should help the state's 1.4 million residents protect the environment.

The project’s price tag now exceeds $100 billion, more than triple the initial estimate.

The ambiguity in medical imaging can present major challenges for clinicians who are trying to identify disease. For instance, in a chest X-ray, pleural effusion, an abnormal buildup of fluid in the lungs, can look very much like pulmonary infiltrates, which are accumulations of pus or blood.

An artificial intelligence model could assist the clinician in X-ray analysis by helping to identify subtle details and boosting the efficiency of the diagnosis process. But because so many possible conditions could be present in one image, the clinician would likely want to consider a set of possibilities, rather than only having one AI prediction to evaluate.

One promising way to produce a set of possibilities, called conformal classification, is convenient because it can be readily implemented on top of an existing machine-learning model. However, it can produce sets that are impractically large. 

MIT researchers have now developed a simple and effective improvement that can reduce the size of prediction sets by up to 30 percent while also making predictions more reliable.

Having a smaller prediction set may help a clinician zero in on the right diagnosis more efficiently, which could improve and streamline treatment for patients. This method could be useful across a range of classification tasks — say, for identifying the species of an animal in an image from a wildlife park — as it provides a smaller but more accurate set of options.

“With fewer classes to consider, the sets of predictions are naturally more informative in that you are choosing between fewer options. In a sense, you are not really sacrificing anything in terms of accuracy for something that is more informative,” says Divya Shanmugam PhD ’24, a postdoc at Cornell Tech who conducted this research while she was an MIT graduate student.

Shanmugam is joined on the paper by Helen Lu ’24; Swami Sankaranarayanan, a former MIT postdoc who is now a research scientist at Lilia Biosciences; and senior author John Guttag, the Dugald C. Jackson Professor of Computer Science and Electrical Engineering at MIT and a member of the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). The research will be presented at the Conference on Computer Vision and Pattern Recognition in June.

Prediction guarantees

AI assistants deployed for high-stakes tasks, like classifying diseases in medical images, are typically designed to produce a probability score along with each prediction so a user can gauge the model’s confidence. For instance, a model might predict that there is a 20 percent chance an image corresponds to a particular diagnosis, like pleurisy.

But it is difficult to trust a model’s predicted confidence because much prior research has shown that these probabilities can be inaccurate. With conformal classification, the model’s prediction is replaced by a set of the most probable diagnoses along with a guarantee that the correct diagnosis is somewhere in the set.

But the inherent uncertainty in AI predictions often causes the model to output sets that are far too large to be useful.

For instance, if a model is classifying an animal in an image as one of 10,000 potential species, it might output a set of 200 predictions so it can offer a strong guarantee.

“That is quite a few classes for someone to sift through to figure out what the right class is,” Shanmugam says.

The technique can also be unreliable because tiny changes to inputs, like slightly rotating an image, can yield entirely different sets of predictions.

To make conformal classification more useful, the researchers applied a technique developed to improve the accuracy of computer vision models called test-time augmentation (TTA).

TTA creates multiple augmentations of a single image in a dataset, perhaps by cropping the image, flipping it, zooming in, etc. Then it applies a computer vision model to each version of the same image and aggregates its predictions.

“In this way, you get multiple predictions from a single example. Aggregating predictions in this way improves predictions in terms of accuracy and robustness,” Shanmugam explains.

Maximizing accuracy

To apply TTA, the researchers hold out some labeled image data used for the conformal classification process. They learn to aggregate the augmentations on these held-out data, automatically augmenting the images in a way that maximizes the accuracy of the underlying model’s predictions.

Then they run conformal classification on the model’s new, TTA-transformed predictions. The conformal classifier outputs a smaller set of probable predictions for the same confidence guarantee.

“Combining test-time augmentation with conformal prediction is simple to implement, effective in practice, and requires no model retraining,” Shanmugam says.

Compared to prior work in conformal prediction across several standard image classification benchmarks, their TTA-augmented method reduced prediction set sizes across experiments, from 10 to 30 percent.

Importantly, the technique achieves this reduction in prediction set size while maintaining the probability guarantee.

The researchers also found that, even though they are sacrificing some labeled data that would normally be used for the conformal classification procedure, TTA boosts accuracy enough to outweigh the cost of losing those data.

“It raises interesting questions about how we used labeled data after model training. The allocation of labeled data between different post-training steps is an important direction for future work,” Shanmugam says.

In the future, the researchers want to validate the effectiveness of such an approach in the context of models that classify text instead of images. To further improve the work, the researchers are also considering ways to reduce the amount of computation required for TTA.

This research is funded, in part, by the Wistrom Corporation.

For policymakers investigating the effective transition of an economy from agriculture to manufacturing and services, there are complex economic, institutional, and practical considerations. “Are certain regions trapped in an under-industrialization state?” asks Tishara Garg, an economics doctoral student at MIT. “If so, can government policy help them escape this trap and transition to an economy characterized by higher levels of industrialization and better-paying jobs?” 

Garg’s research focuses on trade, economic geography, and development. Her studies yielded the paper “Can Industrial Policy Overcome Coordination Failures: Theory and Evidence from Industrial Zones,” which investigates whether economic policy can shift an economy from an undesirable state to a desirable state. 

Garg’s work combines tools from industrial organization and numerical algebraic geometry. Her paper finds that regions in India with state-developed industrial zones are 38 percent more likely to shift from a low to high industrialization state over a 15-year period than those without such zones.  

The kinds of questions uncovered during her studies aren’t easily answered using standard technical and econometric tools, so she’s developing new ones. “One of my study’s main contributions is a methodological framework that draws on ideas from different areas,” she notes. “These tools not only help me study the question I want to answer, but are also general enough to help study a broader set of questions around multiple challenges.”

The new tools she’s developed, along with a willingness to engage with other disciplines, have helped her discover innovative ways to approach these challenges while learning to work with new ones, options she asserts are actively encouraged at an institution like MIT.

“I benefited from having an open mind and learning different things,” she says.

“I was introduced to academia late”

Garg’s journey from Kaithal, India, to MIT wasn’t especially smooth, as societal pressures exerted a powerful influence. “The traditional path for someone like me is to finish school, enter an arranged marriage, and start a family,” she says. “But I was good at school and wanted to do more.” 

Garg, who hails from a background with limited access to information on career development opportunities, took to math early. “I chose business in high school because I planned to become an accountant,” she recalls. “My uncle was an accountant.”

While pursuing the successful completion of a high school business track, she became interested in economics. “I didn’t know much about economics, but I came to enjoy it,” she says. Garg relishes the pursuit of deductive reasoning that begins with a set of assumptions and builds, step by step, toward a well-defined, clear conclusion. She especially enjoys grappling with the arguments she found in textbooks. She continued to study economics as an undergraduate at the University of Delhi, and later earned her master’s from the Indian Statistical Institute. Doctoral study wasn’t an option until she made it one.

“It took me some time to convince my parents,” she says. She spent a year at a hedge fund before applying to economics doctoral programs in the United States and choosing MIT. “I was introduced to academia late,” she notes. “But my heart was being drawn to the academic path.”

Answering ambitious and important questions

Garg, who hadn’t left India before her arrival in Cambridge, Massachusetts, found the transition challenging. “There were new cultural norms, a language barrier, different foods, and no preexisting social network,” she says. Garg relied on friends and MIT faculty for support when she arrived in 2019. 

“When Covid hit, the department looked out for me,” she says. Garg recalls regular check-ins from a faculty advisor and the kind of camaraderie that can grow from shared circumstances, like Covid-related sheltering protocols. A world that forced her to successfully navigate a new and unfamiliar reality helped reshape how she viewed herself. “Support from the community at MIT helped me grow in many ways,” she recalls, “I found my voice here.”

Once she began her studies, one of the major differences Garg found was the diversity of opinions in her field of inquiry. “At MIT, I could speak with students and faculty specializing in trade, development economics, industrial organization, macroeconomics, and more,” she says. “I had limited exposure to many of these subfields before coming to MIT.” 

She quickly found her footing, leaning heavily on both her past successes and the academic habits she developed during her studies in India. “I’m not a passive learner,” she says. “My style is active, critical, and engaged.”

Conducting her research exposes Garg to new ideas. She learned the value of exploring other disciplines’ approaches to problem-solving, which was encouraged and enabled at MIT. 

One of the classes she came to enjoy most was a course in industrial organization taught by Tobias Salz. “I had little familiarity with the material, and it was highly technical — but he taught it in such a clear and intuitive way that I found myself truly enjoying the class, even though it was held during the pandemic,” she recalls. That early experience laid the groundwork for future research. Salz went on to advise her dissertation, helping her engage with work she would build upon.

“Answering ambitious and important questions is what draws me to the work,” Garg says. “I enjoy learning, I enjoy the creative process of bringing different ideas together and MIT's environment has made it easy for me to pick up new things.”

Working with her advisors at MIT helped Garg formalize her research and appreciate the value of uncovering questions and developing approaches to answer them. Professor Abhijit Banerjee, an advisor and Nobel laureate, helped her understand the importance of appreciating different traditions while also staying true to how you think about the problem, she recalls. “That gave me the confidence to pursue the questions in ways that felt most compelling and personal to me,” she says, “even if they didn’t fit neatly into disciplinary boundaries.”

This encouragement, combined with the breadth of perspectives at MIT, pushed her to think creatively about research challenges and to look beyond traditional tools to discover solutions. “MIT’s faculty have helped me improve the way I think and refine my approach to this work,” she says.

Paying it forward

Garg, who will continue her research as a postdoc at Princeton University in the fall and begin her career as a professor at Stanford University in 2026, singles out her network of friends and advisors for special praise.

“From regular check-ins with my advisors to the relationships that help me find balance with my studies, the people at MIT have been invaluable,” she says. 

Garg is especially invested in mentorship opportunities available as a researcher and professor. “I benefited from the network of friends and mentors at MIT and I want to pay it forward — especially for women, and others from backgrounds like mine,” she says.

She cites the work of her advisors, David Atkin and Dave Donaldson — with whom she is also collaborating on research studying incidences of economic distortions — as both major influences on her development and a key reason she’s committed to mentoring others. “They’ve been with me every step of the way,” she says. 

Garg recommends keeping an open mind, above all. “Some of my students didn’t come from a math-heavy background and would restrict themselves or otherwise get discouraged from pursuing theoretical work,” she says. “But I always encouraged them to pursue their interests above all, even if it scared them.” 

The variety of ideas available in her area of inquiry still fascinates Garg, who’s excited about what’s next. “Don’t shy from big questions,” she says. “Explore the big idea.”

With war continuing to disrupt education for millions of Ukrainian high school and college students, many are turning to online resources, including MIT OpenCourseWare, a part of MIT Open Learning offering educational materials from more than 2,500 MIT undergraduate and graduate courses.

For Ukrainian high school senior Sofiia Lipkevych and other students, MIT OpenCourseWare has provided valuable opportunities to take courses in key subject areas. However, while multiple Ukrainian students study English, many do not yet have sufficient command of the language to be able to fully understand and use the often very technical and complex OpenCourseWare content and materials.

“At my school, I saw firsthand how language barriers prevented many Ukrainian students from accessing world-class education,” says Lipkevych.

She was able to address this challenge as a participant in the Ukrainian Leadership and Technology Academy (ULTA), established by Ukrainian MIT students Dima Yanovsky and Andrii Zahorodnii. During summer 2024 at ULTA, Lipkevych worked on a browser extension that translated YouTube videos in real-time. Since MIT OpenCourseWare was a main source of learning materials for students participating in ULTA, she was inspired to translate OpenCourseWare lectures directly and to have this translation widely available on the OpenCourseWare website and YouTube channel. She reached out to Professor Elizabeth Wood, founding director of the MIT Ukraine Program, who connected her with MIT OpenCourseWare Director Curt Newton.

Although there had been some translations of MIT OpenCourseWare’s educational resources available beginning in 2004, these initial translations were conducted manually by several global partners, without the efficiencies of the latest artificial intelligence tools, and over time the programs couldn’t be sustained, and shut down.

“We were thrilled to have this contact with ULTA,” says Newton. “We’ve been missing having a vibrant translation community, and we are excited to have a ‘phase 2’ of translations emerge.”

The ULTA team selected courses to translate based on demand among Ukrainian students, focusing on foundational subjects that are prerequisites for advanced learning — particularly those for which high-quality, Ukrainian-language materials are scarce. Starting with caption translations on videos of lectures, the team has translated the following courses so far: 18.06 (Linear Algebra), 2.003SC (Engineering Dynamics), 5.60 (Thermodynamics & Kinetics), 6.006 (Introduction to Algorithms), and 6.0001 (Introduction to Computer Science and Programming in Python). They also worked directly with Andy Eskenazi, a PhD student in the MIT Department of Aeronautics and Astronautics, to translate 16.002 (How to CAD Almost Anything - Siemens NX Edition).

The ULTA team developed multiple tools to help break language barriers. For MIT OpenCourseWare’s PDF content available through the ULTA program, they created a specialized tool that uses optical character recognition to recognize LaTeX in documents — such as problem sets and other materials — and then used a few large language models to translate them, all while maintaining technical accuracy. The team built a glossary of technical terms used in the courses and their corresponding Ukrainian translations, to help make sure that the wording was correct and consistent. Each translation also undergoes human review to further ensure accuracy and high quality.

For video content, the team initially created a browser extension that can translate YouTube video captions in real-time. They ultimately collaborated with ElevenLabs, implementing their advanced AI dubbing editor that preserves the original speaker's tone, pace, and emotional delivery. The lectures are translated in the ElevenLabs dubbing editor, and then the audio is uploaded to the MIT OpenCourseWare YouTube channel.

The team is currently finalizing the translation of the audio for class 9.13 (The Human Brain), taught by MIT Professor Nancy Kanwisher, which Lipkevych says they selected for its interdisciplinary nature and appeal to a wide variety of learners.

This Ukrainian translation project highlights the transformative potential of the latest translation technologies, building upon a 2023 MIT OpenCourseWare experiment using the Google Aloud AI dubbing prototype on a few courses, including MIT Professor Patrick Winston’s How to Speak. The advanced capabilities of the dubbing editor used in this project are opening up possibilities for a much greater variety of language offerings throughout MIT OpenCourseWare materials.

“I expect that in a few years we’ll look back and see that this was the moment when things shifted for OpenCourseWare to be truly usable for the whole world,” says Newton.

Community-led language translations of MIT OpenCourseWare materials serve as a high-impact example of the power of OpenCourseWare’s Creative Commons licensing, which grants everyone the right to revise materials to suit their particular needs and redistribute those revisions to the world.

While there isn’t currently a way for users of the MIT OpenCourseWare platform to quickly identify which videos are available in which languages, MIT OpenCourseWare is working toward building this capability into its website, as well as expanding its number of offerings in different languages.

“This project represents more than just translation,” says Lipkevych. “We’re enabling thousands of Ukrainians to build skills that will be essential for the country’s eventual reconstruction. We’re also hoping this model of collaboration can be extended to other languages and institutions, creating a template for making high-quality education accessible worldwide.”

Since its founding 19 years ago as a pioneering collaboration with Portuguese universities, research institutions and corporations, the MIT-Portugal Program (MPP) has achieved a slew of successes — from enabling 47 entrepreneurial spinoffs and funding over 220 joint projects between MIT and Portuguese researchers to training a generation of exceptional researchers on both sides of the Atlantic.

In March, with nearly two decades of collaboration under their belts, MIT and the Portuguese Science and Technology Foundation (FCT) signed an agreement that officially launches the program’s next chapter. Running through 2030, MPP’s Phase 4 will support continued exploration of innovative ideas and solutions in fields ranging from artificial intelligence and nanotechnology to climate change — both on the MIT campus and with partners throughout Portugal.  

“One of the advantages of having a program that has gone on so long is that we are pretty well familiar with each other at this point. Over the years, we’ve learned each other’s systems, strengths and weaknesses and we’ve been able to create a synergy that would not have existed if we worked together for a short period of time,” says Douglas Hart, MIT mechanical engineering professor and MPP co-director.

Hart and John Hansman, the T. Wilson Professor of Aeronautics and Astronautics at MIT and MPP co-director, are eager to take the program’s existing research projects further, while adding new areas of focus identified by MIT and FCT. Known as the Fundação para a Ciência e Tecnologia in Portugal, FCT is the national public agency supporting research in science, technology and innovation under Portugal’s Ministry of Education, Science and Innovation.

“Over the past two decades, the partnership with MIT has built a foundation of trust that has fostered collaboration among researchers and the development of projects with significant scientific impact and contributions to the Portuguese economy,” Fernando Alexandre, Portugal’s minister for education, science, and innovation, says. “In this new phase of the partnership, running from 2025 to 2030, we expect even greater ambition and impact — raising Portuguese science and its capacity to transform the economy and improve our society to even higher levels, while helping to address the challenges we face in areas such as climate change and the oceans, digitalization, and space.”

“International collaborations like the MIT-Portugal Program are absolutely vital to MIT’s mission of research, education and service. I’m thrilled to see the program move into its next phase,” says MIT President Sally Kornbluth. “MPP offers our faculty and students opportunities to work in unique research environments where they not only make new findings and learn new methods but also contribute to solving urgent local and global problems. MPP’s work in the realm of ocean science and climate is a prime example of how international partnerships like this can help solve important human problems."

Sharing MIT’s commitment to academic independence and excellence, Kornbluth adds, “the institutions and researchers we partner with through MPP enhance MIT’s ability to achieve its mission, enabling us to pursue the exacting standards of intellectual and creative distinction that make MIT a cradle of innovation and world leader in scientific discovery.”

The epitome of an effective international collaboration, MPP has stayed true to its mission and continued to deliver results here in the U.S. and in Portugal for nearly two decades — prevailing amid myriad shifts in the political, social, and economic landscape. The multifaceted program encompasses an annual research conference and educational summits such as an Innovation Workshop at MIT each June and a Marine Robotics Summer School in the Azores in July, as well as student and faculty exchanges that facilitate collaborative research. During the third phase of the program alone, 59 MIT students and 53 faculty and researchers visited Portugal, and MIT hosted 131 students and 49 faculty and researchers from Portuguese universities and other institutions.

In each roughly five-year phase, MPP researchers focus on a handful of core research areas. For Phase 3, MPP advanced cutting-edge research in four strategic areas: climate science and climate change; Earth systems: oceans to near space; digital transformation in manufacturing; and sustainable cities. Within these broad areas, MIT and FCT researchers worked together on numerous small-scale projects and several large “flagship” ones, including development of Portugal’s CubeSat satellite, a collaboration between MPP and several Portuguese universities and companies that marked the country’s second satellite launch and the first in 30 years.

While work in the Phase 3 fields will continue during Phase 4, researchers will also turn their attention to four more areas: chips/nanotechnology, energy (a previous focus in Phase 2), artificial intelligence, and space.

“We are opening up the aperture for additional collaboration areas,” Hansman says.

In addition to focusing on distinct subject areas, each phase has emphasized the various parts of MPP’s mission to differing degrees. While Phase 3 accentuated collaborative research more than educational exchanges and entrepreneurship, those two aspects will be given more weight under the Phase 4 agreement, Hart said.

“We have approval in Phase 4 to bring a number of Portuguese students over, and our principal investigators will benefit from close collaborations with Portuguese researchers,” he says.

The longevity of MPP and the recent launch of Phase 4 are evidence of the program’s value. The program has played a role in the educational, technological and economic progress Portugal has achieved over the past two decades, as well.  

“The Portugal of today is remarkably stronger than the Portugal of 20 years ago, and many of the places where they are stronger have been impacted by the program,” says Hansman, pointing to sustainable cities and “green” energy, in particular. “We can’t take direct credit, but we’ve been part of Portugal’s journey forward.”

Since MPP began, Hart adds, “Portugal has become much more entrepreneurial. Many, many, many more start-up companies are coming out of Portuguese universities than there used to be.”  

A recent analysis of MPP and FCT’s other U.S. collaborations highlighted a number of positive outcomes. The report noted that collaborations with MIT and other US universities have enhanced Portuguese research capacities and promoted organizational upgrades in the national R&D ecosystem, while providing Portuguese universities and companies with opportunities to engage in complex projects that would have been difficult to undertake on their own.

Regarding MIT in particular, the report found that MPP’s long-term collaboration has spawned the establishment of sustained doctoral programs and pointed to a marked shift within Portugal’s educational ecosystem toward globally aligned standards. MPP, it reported, has facilitated the education of 198 Portuguese PhDs.

Portugal’s universities, students and companies are not alone in benefitting from the research, networks, and economic activity MPP has spawned. MPP also delivers unique value to MIT, as well as to the broader US science and research community. Among the program’s consistent themes over the years, for example, is “joint interest in the Atlantic,” Hansman says.

This summer, Faial Island in the Azores will host MPP’s fifth annual Marine Robotics Summer School, a two-week course open to 12 Portuguese Master’s and first year PhD students and 12 MIT upper-level undergraduates and graduate students. The course, which includes lectures by MIT and Portuguese faculty and other researchers, workshops, labs and hands-on experiences, “is always my favorite,” said Hart.

“I get to work with some of the best researchers in the world there, and some of the top students coming out of Woods Hole Oceanographic Institution, MIT, and Portugal,” he says, adding that some of his previous Marine Robotics Summer School students have come to study at MIT and then gone on to become professors in ocean science.

“So, it’s been exciting to see the growth of students coming out of that program, certainly a positive impact,” Hart says.

MPP provides one-of-a-kind opportunities for ocean research due to the unique marine facilities available in Portugal, including not only open ocean off the Azores but also Lisbon’s deep-water port and a Portuguese Naval facility just south of Lisbon that is available for collaborative research by international scientists. Like MIT, Portuguese universities are also strongly invested in climate change research — a field of study keenly related to ocean systems.

“The international collaboration has allowed us to test and further develop our research prototypes in different aquaculture environments both in the US and in Portugal, while building on the unique expertise of our Portuguese faculty collaborator Dr. Ricardo Calado from the University of Aveiro and our industry collaborators,” says Stefanie Mueller, the TIBCO Career Development Associate Professor in MIT’s departments of Electrical Engineering and Computer Science and Mechanical Engineering and leader of the Human-Computer Interaction Group at the MIT Computer Science and Artificial Intelligence Lab.

Mueller points to the work of MIT mechanical engineering PhD student Charlene Xia, a Marine Robotics Summer School participant, whose research is aimed at developing an economical system to monitor the microbiome of seaweed farms and halt the spread of harmful bacteria associated with ocean warming. In addition to participating in the summer school as a student, Xia returned to the Azores for two subsequent years as a teaching assistant.

“The MIT-Portugal Program has been a key enabler of our research on monitoring the aquatic microbiome for potential disease outbreaks,” Mueller says.

As MPP enters its next phase, Hart and Hansman are optimistic about the program’s continuing success on both sides of the Atlantic and envision broadening its impact going forward.

“I think, at this point, the research is going really well, and we’ve got a lot of connections. I think one of our goals is to expand not the science of the program necessarily, but the groups involved,” Hart says, noting that MPP could have a bigger presence in technical fields such as AI and micro-nano manufacturing, as well as in social sciences and humanities.

“We’d like to involve many more people and new people here at MIT, as well as in Portugal,” he says, “so that we can reach a larger slice of the population.” 

Meta is suing NSO Group, basically claiming that the latter hacks WhatsApp and not just WhatsApp users. We have a procedural ruling:

Under the order, NSO Group is prohibited from presenting evidence about its customers’ identities, implying the targeted WhatsApp users are suspected or actual criminals, or alleging that WhatsApp had insufficient security protections.

[…]

In making her ruling, Northern District of California Judge Phyllis Hamilton said NSO Group undercut its arguments to use evidence about its customers with contradictory statements...

The DOE secretary made a subtle swipe at the continent’s use of renewables, saying that “when you hitch your wagon to the weather, it’s just a risky endeavor.”