The market for carbon credits has struggled to right itself after years of greenwashing allegations.

"We had no warning," said Honolulu's mayor as he toured the damage.

Named Coatlicue, after an Aztec deity, it will be used for climate predictions, energy planning and corruption prevention, among other issues.

Inside a giant autonomous warehouse, hundreds of robots dart down aisles as they collect and distribute items to fulfill a steady stream of customer orders. In this busy environment, even small traffic jams or minor collisions can snowball into massive slowdowns.

To avoid such an avalanche of inefficiencies, researchers from MIT and the tech firm Symbotic developed a new method that automatically keeps a fleet of robots moving smoothly. Their method learns which robots should go first at each moment, based on how congestion is forming, and adapts to prioritize robots that are about to get stuck. In this way, the system can reroute robots in advance to avoid bottlenecks.

The hybrid system utilizes deep reinforcement learning, a powerful artificial intelligence method for solving complex problems, to figure out which robots should be prioritized. Then, a fast and reliable planning algorithm feeds instructions to the robots, enabling them to respond rapidly in constantly changing conditions.

In simulations inspired by actual e-commerce warehouse layouts, this new approach achieved about a 25 percent gain in throughput over other methods. Importantly, the system can quickly adapt to new environments with different quantities of robots or varied warehouse layouts.

“There are a lot of decision-making problems in manufacturing and logistics where companies rely on algorithms designed by human experts. But we have shown that, with the power of deep reinforcement learning, we can achieve super-human performance. This is a very promising approach, because in these giant warehouses even a 2 or 3 percent increase in throughput can have a huge impact,” says Han Zheng, a graduate student in the Laboratory for Information and Decision Systems (LIDS) at MIT and lead author of a paper on this new approach.

Zheng is joined on the paper by Yining Ma, a LIDS postdoc; Brandon Araki and Jingkai Chen of Symbotic; and senior author Cathy Wu, the Class of 1954 Career Development Associate Professor in Civil and Environmental Engineering (CEE) and the Institute for Data, Systems, and Society (IDSS) at MIT, and a member of LIDS. The research appears today in the Journal of Artificial Intelligence Research.

Rerouting robots

Coordinating hundreds of robots in an e-commerce warehouse simultaneously is no easy task.

The problem is especially complicated because the warehouse is a dynamic environment, and robots continually receive new tasks after reaching their goals. They need to be rapidly redirected as they leave and enter the warehouse floor.

Companies often leverage algorithms written by human experts to determine where and when robots should move to maximize the number of packages they can handle.

But if there is congestion or a collision, a firm may have no choice but to shut down the entire warehouse for hours to manually sort the problem out.

“In this setting, we don’t have an exact prediction of the future. We only know what the future might hold, in terms of the packages that come in or the distribution of future orders. The planning system needs to be adaptive to these changes as the warehouse operations go on,” Zheng says.

The MIT researchers achieved this adaptability using machine learning. They began by designing a neural network model to take observations of the warehouse environment and decide how to prioritize the robots. They train this model using deep reinforcement learning, a trial-and-error method in which the model learns to control robots in simulations that mimic actual warehouses. The model is rewarded for making decisions that increase overall throughput while avoiding conflicts.

Over time, the neural network learns to coordinate many robots efficiently.

“By interacting with simulations inspired by real warehouse layouts, our system receives feedback that we use to make its decision-making more intelligent. The trained neural network can then adapt to warehouses with different layouts,” Zheng explains.

It is designed to capture the long-term constraints and obstacles in each robot’s path, while also considering dynamic interactions between robots as they move through the warehouse.

By predicting current and future robot interactions, the model plans to avoid congestion before it happens.

After the neural network decides which robots should receive priority, the system employs a tried-and-true planning algorithm to tell each robot how to move from one point to another. This efficient algorithm helps the robots react quickly in the changing warehouse environment.

This combination of methods is key.

“This hybrid approach builds on my group’s work on how to achieve the best of both worlds between machine learning and classical optimization methods. Pure machine-learning methods still struggle to solve complex optimization problems, and yet it is extremely time- and labor-intensive for human experts to design effective methods. But together, using expert-designed methods the right way can tremendously simplify the machine learning task,” says Wu.

Overcoming complexity

Once the researchers trained the neural network, they tested the system in simulated warehouses that were different than those it had seen during training. Since industrial simulations were too inefficient for this complex problem, the researchers designed their own environments to mimic what happens in actual warehouses.

On average, their hybrid learning-based approach achieved 25 percent greater throughput than traditional algorithms as well as a random search method, in terms of number of packages delivered per robot. Their approach could also generate feasible robot path plans that overcame congestion caused by traditional methods.

“Especially when the density of robots in the warehouse goes up, the complexity scales exponentially, and these traditional methods quickly start to break down. In these environments, our method is much more efficient,” Zheng says.

While their system is still far away from real-world deployment, these demonstrations highlight the feasibility and benefits of using a machine learning-guided approach in warehouse automation.

In the future, the researchers want to include task assignments in the problem formulation, since determining which robot will complete each task impacts congestion. They also plan to scale up their system to larger warehouses with thousands of robots.

This research was funded by Symbotic.

Like many people who end up going into physics, Sophia Henneberg had a hard time, when she was young, choosing between that discipline and mathematics. Both subjects came easily to her, and she — unlike many of her peers — thought they were fun. Henneberg grew up in a small town in central Germany, and it was not until one week before applying to college that she decided on physics, reasoning that it would still give her the chance to do plenty of math, while also affording opportunities to connect with a broad range of applications. 

Midway through her undergraduate studies at Goethe University in Frankfurt, she started taking courses in plasma physics and almost instantly knew that she had found her niche. “Most of the visible material in the universe is in the form of hot, ionized gas called plasma, so studying that is really fundamental,” she says. “And there’s this amazing application, fusion, which has the potential to become an unlimited energy source.”

Early on, Henneberg resolved to try to make that potential a reality, and she’s been pursuing that goal at MIT since becoming the Norman Rasmussen Career Development Assistant Professor in the Department of Nuclear Science and Engineering in fall of 2025. Her research focus is on stellarators — a kind of fusion machine that has been overshadowed for many decades by another fusion device called the tokamak.  Both of these machines rely on magnetic confinement — using powerful magnetic fields to compress a plasma into a tiny volume causing some of the atoms within this dense cluster to fuse together, unleashing energy in the process. In the tokamak, the plasma assumes the shape of a donut. In a stellarator, the plasma is also contained within a rounded loop, only this one resembles a twisted donut.

As a PhD candidate at the University of York (in the United Kingdom), Henneberg studied the instabilities that can arise in tokamaks, where plasma temperatures often exceed 100 million degrees Celsius and currents induced within the plasma can attain speeds of roughly 100 kilometers per second. In such an ultra-extreme setting — more than six times hotter than the core of the sun — sudden surges of energy, leading to something akin to small-scale solar flares, can breach the magnetic cage enclosing the plasma, thereby disrupting the fusion process and possibly damaging the reactor itself. Henneberg started hearing about stellarators in her classes and, after a bit of research, she came to realize that “they could be much more stable if you design them in the right way.”

Striking a favorable balance

In 2016, she began a postdoctoral fellowship at the Max Planck Institute (MPI) for Plasma Physics in Greifswald, Germany, joining the Stellarator Theory Group. Greifswald may well have been the best place for her to carry out stellarator research, given that the world’s biggest and most advanced reactor of this type, Wendelstein 7-X (W7-X), was based there, and experiments were just starting in the year she arrived.

Her main assignment at MPI was to work on stellarator optimization, figuring out the best way to design the reactor to meet the engineering and physics goals — a task not unlike that of tuning a car to achieve maximum fuel efficiency or, for a racecar, maximum speed. Henneberg’s interest in optimization continues to this day, remaining central to her research agenda at MIT.

“If you want to design a stellarator, there are two principal components you can look at,” she says. The first relates to the shape of the boundary, or cage, into which the plasma will ultimately be confined. This shape is constrained by magnetic fields that are generated, in turn, by a series of superconducting coils that might range in number anywhere from around 4 to 50. In stellarators, the coils tend to be bent rather than circular. That gives rise to twists in the magnetic fields, but it also makes the coils more complicated and likely more expensive. Henneberg has come up with ways to simplify the optimization process — one of which involves designing the plasma boundary and the shape of the coils in the same step rather than looking at them separately. 

“We’ve now reached the point where stellarator performances can exceed those of tokamaks, because we’re able to optimize them very well, but you have to put the effort in,” she says. “You can’t get good performance out of just any twisty donut.”

The best of both worlds

In a 2024 paper, Henneberg and her former Greifswald colleague, Gabriel Plunk, introduced the notion of a stellarator-tokamak hybrid reactor. The goal, they wrote, is both “simple and compelling: to combine the strengths of the two concepts into a single device” that outperforms either of the existing modes.

One of Henneberg’s major preoccupations at present is exploring ways of converting a tokamak into a stellarator that basically entails adding just a few coils — of the bent variety — that can be turned on or off. “This can be an easy way for people in the tokamak community to think more about the possible benefits of the stellarator,” she says. While no one has yet built a hybrid, at least one university has secured funding to do so.

Interest in stellarators has been steadily mounting in recent years, a fact that delights Henneberg. When she started working in this area almost a decade ago, the field of stellarator optimization was tiny and there were very few people she could converse with. There’s much more research going on today, which means that more ideas are coming out, along with some exciting results. Commercial interest is growing as well, and Henneberg has been in contact with several stellarator startup companies, including Type One Energy and Thea Energy in the United States and Proxima Fusion and Gauss Fusion in Germany.

“It seems to me that most new startups these days are focusing on stellarators,” Henneberg says. “With so many companies now entering the field, it can seem like the technical issues involved in fusion are already solved, but there are still many interesting open questions. I’m working on improved designs that advance both the physics and the economic feasibility.”

That’s where her students come in. She believes that one part of her role as an MIT professor is to train the next generation of stellarator experts — people who will help, for instance, to design effective coils that are easy to make, as well as to improve reactor performance overall. 

During her first term, she co-taught the renowned Fusion Design (22.63) course alongside MIT Professor Dennis Whyte. This course has had a remarkable influence on the fusion community, leading to nine published papers with over 1,000 citations and inspiring the creation of several companies. In the fall 2025 version of this course, students were charged with comparing designs for stellarators with machines that relied on a different way of confining the plasma called magnetic mirrors.

After just a few months at MIT, Henneberg has been impressed with her students, calling them “highly motivated and a lot of fun to work with.” She’s confident that her research group will soon be making progress.

She is also happy to be affiliated with MIT’s Plasma Science and Fusion Center, which is internationally recognized as a leading university laboratory in this field. “It’s great to have so many experts [primarily in tokamaks] in one place that I can work with and learn from,” Henneberg says. “Because of my interest in hybrid reactors, my research will really benefit from all the expertise here on the tokamak side.”

Each spring, river herring populations migrate from Massachusetts coastal waters to begin their annual journey up rivers and streams to freshwater spawning habitat. River herring have faced severe population declines over the past several decades, and their migration is extensively monitored across the region, primarily through traditional visual counting and volunteer-based programs. 

Monitoring fish movement and understanding population dynamics are essential for informing conservation efforts and supporting fisheries management. With the annual herring run getting underway this month, researchers and resource managers once again take on the challenge of counting and estimating the migrating fish population as accurately as possible. 

A team of researchers from the Woodwell Climate Research Center, MIT Sea Grant, the MIT Computer Science and Artificial Intelligence Lab (CSAIL), MIT Lincoln Laboratory, and Intuit explored a new monitoring method using underwater video and computer vision to supplement citizen science efforts. The researchers — Zhongqi Chen and Linda Deegan from the Woodwell Climate Research Center, Robert Vincent and Kevin Bennett from MIT Sea Grant, Sara Beery and Timm Haucke from MIT CSAIL, Austin Powell from Intuit, and Lydia Zuehsow from MIT Lincoln Laboratory — published a paper describing this work in the journal Remote Sensing in Ecology and Conservation this February. 

The open-access paper, “From snapshots to continuous estimates: Augmenting citizen science with computer vision for fish monitoring,” outlines how recent advancements in computer vision and deep learning, from object detection and tracking to species classification, offer promising real-world solutions for automating fish counting with improved efficiency and data quality. 

Traditional monitoring methods are constrained by time, environmental conditions, and labor intensity. Volunteer visual counts are limited to brief daytime sampling windows, missing nighttime movement and short migration pulses, when hundreds of fish pass by within the span of a few minutes. While technologies like passive acoustic monitoring and imaging sonar have advanced continuous fish monitoring under certain conditions, the most promising and low-cost option — manual review of underwater video — is still labor-intensive and time-consuming. With the growing demand for automated video processing solutions, this study presents a scalable, cost-effective, and efficient deep learning-based system for reliable automated fish monitoring. 

The team built an end-to-end pipeline — from in-field underwater cameras to video labeling and model training — to achieve automated, computer vision-powered fish counting. Videos were collected from three rivers in Massachusetts: the Coonamessett River in Falmouth, the Ipswich River (Ipswich), and the Santuit River in Mashpee. 

To prepare the training dataset, the team selected video clips with variations in lighting, water clarity, fish species and density, time of day, and season to ensure that the computer vision model would work reliably across diverse real-world scenarios. They used an open-source web platform to manually label the videos frame-by-frame with bounding boxes to track fish movement. In total, they labeled 1,435 video clips and annotated 59,850 frames. 

The researchers compared and validated the computer vision counts with human video reviews, stream-side visual counts, and data from passive integrated transponder (PIT) tagging. They concluded that models trained on diverse multi-site and multi-year data performed best and produced season-long, high-resolution counts consistent with traditionally established estimates. Going one step further, the system provided insights into migration behavior, timing, and movement patterns linked to environmental factors. Using video from the 2024 Coonamesset River migration, the system counted 42,510 river herring and revealed that upstream migration peaked at dawn, while downstream migration was largely nocturnal, with fish utilizing darker, quieter periods to avoid predators.

With this real-world application, the researchers aim to advance computer vision in fisheries management and provide a framework and best practices for integrating the technology into conservation efforts for a wide range of aquatic species. “MIT Sea Grant has been funding work on this topic for some time now, and this excellent work by Zhongqi Chen and colleagues will advance fisheries monitoring capabilities and improve fish population assessments for fisheries managers and conservation groups,” Vincent says. “It will also provide education and training for students, the public, and citizen science groups in support of the ecologically and culturally important river herring populations along our coasts.”

Still, continued traditional monitoring is essential for maintaining consistency in long-term datasets until fisheries management agencies fully implement automated counting systems. Even then, computer vision and citizen science should be seen as complementary. Volunteers will be necessary for camera maintenance and for contributing directly to the computer vision workflow, from video annotation to model verification. The researchers envision that integrating citizen observations and computer vision-generated data will help create a more comprehensive and holistic approach to environmental monitoring.

This work was funded by MIT Sea Grant, with additional support provided by the Northeast Climate Adaptation Science Center, an MIT Abdul Latif Jameel Water and Food Systems seed grant, the AI and Biodiversity Change Global Center (supported by the National Science Foundation and the Natural Sciences and Engineering Research Council of Canada), and the MIT Undergraduate Research Opportunities Program.

Little Is Known About AI That Could Affect Millions of Seniors' Care

SAN FRANCISCO – The Electronic Frontier Foundation (EFF) today filed a Freedom of Information Act (FOIA) lawsuit against the Centers for Medicare & Medicaid Services (CMS) seeking records about a multi-state program that is using AI to evaluate requests for medical care.

"Tasking an algorithm with making determinations about treatment can create unwarranted—and even discriminatory—delays or denials of necessary medical care," said Kit Walsh, EFF’s Director of AI and Access-to-Knowledge Legal Projects. "Given these serious risks, the public requires transparency that it hasn't gotten. We're suing to get badly needed answers about how Medicare's AI experiment works."

Announced by CMS Administrator Dr. Mehmet Oz last year, the pilot program known as WISeR (Wasteful and Inappropriate Service Reduction) uses AI to assess prior authorization requests from Medicare beneficiaries. Previously rare in original Medicare, prior authorization requires medical providers to obtain advance approval from a patient’s health insurer before delivering certain treatments or services as a condition of coverage.

Unfortunately, there is little information about how the AI algorithms used in WISeR work, including what training data they rely on. It remains unclear whether WISeR has any safeguards against systemic flaws such as algorithmic bias, privacy violations, and wrongful denials of care.

Healthcare experts, care providers, and lawmakers have all raised alarms that WISeR may cause serious harm to patients by relying on AI unless it has the necessary safeguards. Despite this widespread criticism, WISeR was rolled out in six states in January, potentially affecting as many as 6.4 million Medicare beneficiaries, according to one estimate.

By design, WISeR incentivizes contracted companies to deny prior approval against the best interests of patients. Vendors are compensated, in part, on the volume of healthcare services they deny and are entitled to as much as 20 percent of the associated savings. Just weeks after WISeR's launch, hospitals and other health care providers started reporting delays in care approval, communication gaps, and administrative strain.

Earlier this year, EFF submitted a FOIA request to CMS asking for records related to WISeR. Among other records, the request sought agreements with software vendors participating in WISeR; records related to any tests for accuracy, bias, or hallucinations in vendors' technology; and records related to any audits, monitoring, or evaluation of WISeR and participating vendors. To date, CMS has not provided any of these records to EFF. EFF's FOIA lawsuit asks for their immediate processing and release.

"The public has a right to know more about the algorithms driving decisions around their healthcare," said Tori Noble, Staff Attorney at EFF. "Without greater transparency, patients, providers, and policymakers will continue to be left in the dark.”

EFF thanks Stanford Law School's Juelsgaard Intellectual Property & Innovation Clinic for their help in preparing this lawsuit.

For the complaint: https://www.eff.org/document/complaint-eff-v-cms-medicare-wiser-foia

Tags: transparencyartificial intelligence and machine learningautomated decision-makingContact:  ToriNobleStaff Attorneytori@eff.org

Batteries that use solid metal as their charge-carrying electrolyte could potentially be a safer and far more energy-dense alternative to lithium-ion batteries. However, these solid-state batteries have been plagued by the formation of metallic cracks called dendrites that cause them to short circuit.

The problem has so far prevented such batteries from becoming a major player in energy storage. But now, research from MIT could finally help engineers find a way to get past this hurdle.

For decades, many researchers have treated dendrites as largely the result of mechanical stress — like cracks that form on the sidewalk when a tree root grows underneath. But MIT engineers have discovered the exact opposite: Faster dendrite growth was associated with lower stress levels in a commonly used battery electrolyte material. Using a new technique that allowed them to directly measure the stress around growing dendrites, the researchers found cracks formed at stress levels as low as 25 percent of what would be expected under mechanical stress alone.

The experiments, published in Nature today, instead revealed another culprit: chemical reactions caused by high electrical currents that weaken the electrolyte and make it more susceptible to dendrite growth. Researchers had previously proposed that such reactions cause dendrite growth, but the new study provides the first experimental data on the interplay between chemical and mechanical stress in dendrite formation.

“Direct measurement techniques allowed us to see how tough the material is as we cycle the cell,” says Cole Fincher, the paper’s first author and an MIT PhD student in materials science and engineering. “What we saw was that if you just test the ceramic electrolyte on the benchtop, it’s about as tough as your tooth. But during charging, it gets a lot weaker — closer to the brittleness of a lollipop.”

The findings reveal why developing stronger electrolytes alone hasn’t solved the decades-old dendrite problem. It also points to the importance of developing more chemically stable materials to finally fulfill the promise of high-density solid-state batteries.

“There’s a large community of researchers that are constantly trying to discover and design better solid electrolytes to enable the solid-state battery,” says senior author Yet-Ming Chiang, MIT’s Kyocera Professor of Materials Science and Engineering. “This study provides guidance in those efforts. We discovered a new mechanism by which these dendrites grow, allowing us to explore ways to design around it to make solid-state batteries successful.”

Joining Fincher and Chiang on the paper are MIT PhD student Colin Gilgenbach; Thermo Fisher Scientific scientists Christian Roach and Rachel Osmundsen; MIT.nano researcher Aubrey Penn; MIT Toyota Professor in Materials Processing W. Craig Carter; MIT Kyocera Professor of Materials Science and Engineering James LeBeau; University of Michigan Professor Michael Thouless; and Brown University Professor Brian W. Sheldon.

Measuring stress

Dendrites have presented a major roadblock to battery development since the 1970s. One reason lithium-ion batteries have become ubiquitous while other approaches have stalled is that their commonly used graphite anodes are less susceptible to dendrite formation. That’s a shame because solid-state batteries that use lithium metal as an anode and a solid electrolyte could theoretically store far more energy in the same sized package with less weight. They could thus enable longer-lasting phones and laptops, or electric cars with double the range of today’s options.

“There’s no more energy-dense form of lithium than lithium metal,” Chiang says. “But the dendrite problem has limited progress with solid-state batteries.”

Lithium metal is soft like taffy. Fincher, who has been studying the dendrite problem in the labs of Chiang and Carter, says one puzzle is how such a soft material can penetrate into the hard electrolyte materials being explored for use in solid-state batteries.

“The ceramics that have been used in these applications are stiff, like a coffee mug, so it’s been hoped that solid-state batteries would stop this relatively soft dendrite from growing,” Fincher explains.

Believing that mechanical stress causes dendrites, scientists have worked to develop stronger electrolytes that can withstand more mechanical stress. Some researchers have proposed that chemical reactions play a role in dendrite formation, but how those reactions worked with mechanical stress was not known.

For their Nature study, the researchers set out to directly observe mechanical and chemical changes in a commonly used solid-state electrolyte material as dendrites grew. Solid-state batteries are typically organized like a sandwich, which makes it hard to look inside the middle electrolyte layer. For their first experiment, the researchers developed a special solid-state battery cell in which the ceramic layers can be observed from the side, allowing the researchers to watch dendrite growth occurring in the electrolyte.

The researchers also used a measurement technique called birefringence microscopy to precisely measure the stress around the dendrite, which Fincher developed as part of his PhD thesis.

“It works the same way as polarized sunglasses when you look at something like a windshield,” Fincher explains of the technique. “When light comes through, residual stresses in the glass enable light of some orientations to pass faster than others, and that can give rise to observable rainbow patterns. These patterns can be used to measure stress.”

The technique gave the researchers a way to both visualize and quantify stress around actively growing dendrites for the first time, leading to the unexpected findings.

“Normally you would expect that the faster a dendrite grows, the more stress it creates,” Chiang says. “Instead, we observed exactly the opposite. The faster it grew, the lower the stress around it, meaning the solid electrolyte is breaking under a lower stress, and therefore it’s been embrittled.”

In fact, the dendrites grew at stress levels far weaker than expected. Fincher describes the weaker electrolyte as electrochemically corroded.

“Imagine you test a piece of glass one day, and the next day it’s only a quarter as strong,” Chiang says. “It was very surprising.”

Led by LeBeau, the researchers then cooled the electrolyte to extremely low temperatures and applied a powerful imaging technique called cryogenic scanning transmission electron microscopy that allowed them to study the area around the dendrite on nearly atomic scales. The imaging revealed that the passage of ionic current through the material had caused chemical reactions that made it more brittle.

“The electric current drives the flow of lithium ions through the solid electrolyte,” Chiang explains. “That causes a highly concentrated flow of lithium ions at the dendrite tip. We believe that leads to a chemical reduction of the material compound, which leads to its decomposition into new phases. You start with a crystalline phase of the electrolyte, then there’s a volume contraction after the deposition that is consistent with the embrittlement we see.”

Toward better batteries

The experiment was done on one of the most stable electrolytes used in solid-state batteries, making the researchers confident the findings will carry over to other electrolyte materials.

“This tells us we have to look for electrolyte materials that are even more stable, especially when in contact with lithium metal, which chemically speaking is very reducing,” Chiang says. “This will help direct the search for new materials.”

For instance, Chiang says now that they understand more about the chemical changes causing embrittlement, researchers could explore materials that actually get tougher as cracks grow.

The researchers say it will take more work to figure out what electrochemical reactions are taking place to make the electrolyte so much weaker. But they say their approach for directly observing stresses could also help improve materials for use in devices like fuel cells and electrolyzers.

The work was supported by the center for Mechano-Chemical Understanding of Solid Ionic Conductors, a Department of Energy Engineering Frontiers Research Center, the National Science Foundation, and Fincher’s Department of Defense Science and Engineering Graduate Fellowship, and was carried out using MIT.nano facilities.

After years of tech industry experiments, smartglasses with embedded cameras and microphones have finally gone mainstream. And, disturbingly, sometimes it's not just their owners who are watching what these devices record. In this week's EFFector newsletter, we're taking a closer look at the privacy implications of Meta Ray-Bans, and sharing all the latest in the fight for privacy and free speech online.

JOIN OUR NEWSLETTER

For over 35 years, EFFector has been your guide to understanding the intersection of technology, civil liberties, and the law. This week's issue covers EFF's new executive director; how publishers blocking the Internet Archive threaten the web's historical record; and why you should think twice before buying or using Meta’s Ray-Bans.

Prefer to listen in? EFFector is now available on all major podcast platforms. This week, we're chatting with EFF Security and Privacy Activist Thorin Klosowski about smartglasses and privacy. And don't miss the EFFector news quiz. You can find the episode and subscribe on your podcast platform of choice: 

%3Ciframe%20height%3D%22200px%22%20width%3D%22100%25%22%20frameborder%3D%22no%22%20scrolling%3D%22no%22%20seamless%3D%22%22%20src%3D%22https%3A%2F%2Fplayer.simplecast.com%2Fc139744a-aad2-4d31-8b5e-84764a13bf2f%3Fdark%3Dfalse%22%20allow%3D%22autoplay%22%3E%3C%2Fiframe%3E Privacy info. This embed will serve content from simplecast.com

   

Want to stay in the fight for privacy and free speech online? Sign up for EFF's EFFector newsletter for updates, ways to take action, and new merch drops. You can also fuel the fight against online surveillance when you support EFF today!

Interviewer: Jillian York

Jacob Mchangama is a Danish lawyer, human-rights advocate, and public commentator. He is the founder and director of Justitia, a Copenhagen-based think tank focusing on human rights, freedom of speech, and the rule of law. His new book with Jeff Kosseff, The Future of Free Speech: Reversing the Global Decline of Democracy's Most Essential Freedom, comes out on April 7th.

Jillian York: Welcome, Jacob. I'm just going to kick off with a question that I ask everyone, which is: what does free speech mean to you?

Jacob Mchangama: I like to use the definition that Spinoza, the famous Dutch renegade philosopher, used. He said something along the lines, and I'm paraphrasing here, that free speech is the right of everyone to think what they want and say what they think, or the freedom to think what they want and say what they think. I think that's a pretty neat definition, even though it may not be fully exhaustive from sort of a legal perspective, I like that. 

JY: Excellent. I really like that. I'd like to know what personally shaped your views and also what brought you to doing this work for a living. 

JM: I was born in Copenhagen, Denmark, which is a very liberal, progressive, secular country. And for most of my youth and sort of young adulthood, I did not think much about free speech. It was like breathing the air. It was essentially a value that had already been won. This was up until sort of the mid-naughties. I think everyone was sort of surfing the wave of optimism about freedom and democracy at that time. 

And then Denmark became sort of the epicenter of a global battle of values over religion, the relationship between free speech and religion with the whole cartoon affair. And that's really what I think made me think deep and hard about that, that suddenly people were willing to respond to cartoonists using crayons with AK-47s and killings, but also that a lot of people within Denmark suddenly said, “Well, maybe free speech doesn't include the right to offend, and maybe you're punching down on a vulnerable minority,” which I found to be quite an unpersuasive argument for restricting free speech. 

But what's also interesting was that you saw sort of how positions on free speech shifted. So initially, people on the left were quite apprehensive about free speech because they perceived it to be about an attack on minorities, in this case, Muslim immigrants in Denmark. Then the center right government came into power in Denmark, and then the narrative quickly became, well, we need to restrict certain rights of hate preachers and others in order to defend freedom and democracy. And then suddenly, people on the right who had been free speech absolutists during the cartoon affair were willing to compromise on it, and people on the left who had been sort of, well, “maybe free speech has been taken too far” were suddenly adamant that this was going way too far, and unfortunately, that is very much with us to this day. It's difficult to find a principled, consistent constituency for free speech. 

JY: That's a great way of putting it. I feel like, with obvious differences from country to country, it feels like that kind of polarization is true everywhere, including the bit about flipping sides. I guess my next question, then, is: what do you feel like most people get wrong about free speech?

JM: I think there's a tendency—and I'm talking especially in the West, in the traditional free and open democracies—I think there's a huge tendency to take all the benefits of free speech for granted and focus myopically on the harms, real and perceived, of speech. I mean, just the fact that you and I can sit here, you know, I don't know where you are in the world, but you and I can have a direct, live, uncensored conversation…that is something that you know was unimaginable not that long ago, and we just take that for granted. We take it for granted that we can have access to all the information in the world that would previously have required someone to spend years in libraries, traveling the world, finding rare manuscripts.

We take it for granted, but this is the difference between us and say dissidents in Iran or Russia or Venezuela. We take it for granted that we can go online and vent against our governments and say things, and we can also vent things on social issues that might be deeply offensive to other people, but generally we don't face the risk of being imprisoned or tortured. But that's just not the case in many other countries. 

So, I think those benefits, and also, I would say, when you look at the historical angle, every persecuted or discriminated against group that has sought and achieved a higher degree of equal dignity, equal protection under the law, has relied on speech. First they relied on speech, then they could rely on free speech at some point, but initially they didn't have free speech right? So whether it's abolitionist the civil rights movement in the United States, you know my good friend Jonathan Rauch, who was sort of at the forefront of of securing same sex marriage in the United States, knows that was a fight that very much relied on speech. And women's rights…fierce women, who would protest outside the White House and burn in effigy figures of the President, would go to prison. Women didn't have political power. They didn't have guns. They didn't have economic power, they had speech, and that's what you need, to petition the government, to shine a light on abuse, to rally other allies and so on. And I think unfortunately, we've unlearned those hugely important precedents for why we have free speech today. 

JY: I’m definitely going to come back to that. But first I want to ask you about the new book you have coming out with Jeff Kosseff, The Future of Free Speech: Reversing the Global Decline of Democracy's Most Essential Freedom. I'm very excited, I’ve pre-ordered it. 

So, in light of that, I’ve got a two part question: First, what are some of the trends that concern you the most about what’s going on today? And then, what do you think we need to do to ensure that there is a future for free speech?

JM: So first of all, I was thrilled to be able to write it with Jeff, because Jeff is such an authority on First Amendment section 230 issues. But from the personal perspective, you could say that this book sort of continues where my previous book on the history of free speech finishes.

And so, based on the idea that we are living through a free speech recession that has become particularly acute in this digital age, where we see what I term as various waves of elite panic that lead to attempts to impose sort of top down controls on online speech in particular—and this is not only in the countries where you'd expect it, like China and Russia and Iran, but increasingly also in open democracies that used to be the heartland of free speech—there's a tendency, I think, in democracies, to view free speech no longer as sort of a competitive advantage against authoritarian states, or a right that would undermine authoritarians, but as sort of a Trojan horse which allows the enemies of democracies, both at home and abroad, to weaponize free speech against democracy, and so that's why the overwhelming

legislative initiatives and framing of free speech is often “this is a danger.” This is something we need to do something about. We need to do something about disinformation. We need to do something about hate speech. We need to do something about extremism. We need to do something about, you know, we need to have child safety laws. We need age verification. And you know, you know the list all too well. 

JY: I do, absolutely.

JM: Where I think where free speech advocates often fall short, is that we're very good at sort of talking about the slippery slope and John Stuart Mill and all these things, and that's important, but very often we don't have compelling proposals to sell to people who are not sort of civil libertarians at heart, and who are generally in favor of free speech, but who are frightened about particular developments at particular manifestations of speech that they think have become so dangerous to you know, freedom, democracy, whatever interest that they're willing to compromise free speech. 

And so we try to point to some concrete examples of—giving life to the old cliché—fighting bad speech with better speech. So some of those examples are counter speech. There are some great examples. One of them is from Brazil, where there was a black weather woman who was the first black weather woman to be sort of on a prominent TV channel, and she was met with brutal racism. So, you know, what should have been a happy moment for her became quite devastating. And so there was this NGO that printed billboards of these very nasty racist comments, blurred the identity of the user who had said it, but then put them in the neighborhoods where these people lived. So that was a very powerful way to confront Brazilians with the fact that, you know, racism is alive. It's right here in your neighborhood. And you know they used the N word and everything, and nothing was censored in terms of this racism, which was put right in front of it of everyone, and it actually led to a lot of people sort of deleting their comments and someone apologizing, and led to, I think, a fruitful debate in Brazilian society. 

Then you have other types of counter speech. One of them is a Swedish journalist called Mina Dennert. She started the “I am here” movement. So it's a counter speech movement, which I think spans 150,000 volunteers across 15 countries. And they use counter speech online, typically on Meta platforms, I think, where they essentially gather together and push back against hate speech, not necessarily to convince the speaker that they're wrong, but to give support to those who are the victims, but also to essentially convince what is often termed the movable middle, to show them that there are people who disagree with racist hate speech, and there's actually empirical data to suggest that these can be effective strategies. You can also use humor. 

Daryl Davis is a very extreme example. He's a black jazz musician who has made it his life mission to befriend members of the KKK. And he has converted around 200 members of the KKK, to essentially leave it and he does that by just having a conversation. Because if your worldview is that blacks are inferior and should not enjoy equal rights, and you have a conversation with someone in a way where it becomes impossible for you to uphold that worldview, because the person in front of you is clearly someone who's intelligent, articulate, who can counter all your your preconceived notions, then it becomes very difficult to uphold that worldview right? And you can imagine that those members who leave the KKK then become agents of change within their former communities. 

So there are various counter speech strategies that have shown a promise, and at the Future of Free Speech [think tank] that I direct, we've developed these toolkits, and we do teachings around the world, I think we've translated them into nine or ten languages. So it's not a panacea, obviously, to everything that's going on, but it's something quite practical, I think. And the good thing about it is also that it doesn't depend on an official definition of hate speech. If you're concerned about a particular type of speech, you can use counter speech to counter it. But you're not engaging in censorship, and we don't have to agree on what the definition of hate speech is. In that way, it’s hopefully an empowering tool. 

And another example: we talk about how Taiwan has been quite an inspiring case for using crowd sourced fact checking, for using sort of a bottom up approach to fighting disinformation from China, but also around Covid, so zero lockdowns and no centralized censorship, and they’re doing better than a lot of Western democracies that use more illiberal methods and the crowd sourced fact checking pioneered in Taiwan is what inspired Bird Watch on Twitter prior to its being taking over by Elon Musk, and which is now community notes on X, which I actually think for all the things you might dislike about X, is a feature that is quite promising. 

JY: Definitely.  I absolutely agree with that, and I'm really glad you mentioned your previous book, which I loved, and the idea of a free speech recession. 

You’ve done so much of this work all over the world, and have learned from people in different places and tried to understand the challenges they’re facing in terms of free speech. We actually started this project, Speaking Freely, primarily to share those different perspectives and to bring them to our readership, the majority of which comes from the U.S. What I’d like to ask you, then, is what do you feel that we in the “West” or in more open societies have to learn from free speech activists in the rest of the world?

JM: Just…the bravery of say, Iranians who now face complete—and this was even before the attacks by the US and Israel—complete internet bans. But who have also relied on social media platforms and digital creativity to circumvent official propaganda and censorship. I think those types of societies provide sort of a real time experiment, right? You know, okay, we have we have social media, and it's messy, and sometimes it's ugly, and sometimes some of these tech companies do things that we disapprove of, but you know the cure in terms of further government control, for instance, let's say, getting rid of section 230, adding age verification laws, trying to create exceptions to the First Amendment in cyberspace…we have societies where that is happening, albeit, of course, at a very extreme scale. But would you really trade the freedoms, however messy they are, for that kind of society? 

And then, I also worry a lot about the state of affairs in Europe, where I'm from, where it's not unusual if you're in Germany, to have the police show up at your door if you've insulted a powerful politician. For the book, I interviewed an Israeli, Jewish woman who lives in Berlin. She's on the far left and very opposed to to Israel's policies, and she's been arrested four times for for protesting with a plaque that says, “as an Israeli Jew, stop the genocide in Gaza.” And again, you can agree or disagree whether there's a genocide, but that's just political speech. Yet the optics of a Jew—an Israeli, Jewish woman—being arrested by German police in Berlin in the name of fighting antisemitism is, I think, absurd, right? 

JY: I’m laughing only because I think I’ve said that exact sentence in an interview with the German press.

JM: But this is the reality right now. And I think it's also a good example of the fact that there have been people on the left in Europe who have said, well, we need to do something about the far right. And therefore it's okay to crack down, you know, use hate speech laws and so on. And then October 7 happened, and suddenly you see a lot of minorities and people on the left who are becoming the targets of laws against hate speech or glorification of terrorism and so on and so forth. And I think that's a powerful case for why you want a pretty hard nosed principle of consistent protection of free speech, also online. And, given the priorities of the current administration in the United States, I think that if the First Amendment and section 230 were not in place in the United States, the kind of laws that you have in Europe would be very moldable for the current administration to go after. I mean, it’s already going after its enemies, real and perceived, but it often loses in court exactly because of constitutional protections, including the First Amendment. But if that protection wasn't there, they would be much more successful, I think, in going after speech that they don't like.

JY: That’s such a fantastic answer, and I’m in total agreement. I was actually living in Berlin until quite recently and saw quite a bit of that firsthand. It’s really troubling. 

I want to shift course for a moment. We hopefully have some young people reading this as well, and I think right now in this moment where age verification proposals are happening everywhere—which we at EFF are really concerned about—it’s important to speak to them as well. What advice would you give to young readers who are coming of age around the topic of free speech and who are interested in doing this sort of work?

JM: I think young people are obviously immersed in the digital age, and some of them may never have opened a physical book. I don't know. Maybe it's a Boomer prejudice when I say that, but, but, I don't think it's a stretch to imagine that the vast majority of speech and expression that they're confronted with is through devices of a sort. I think it's crucial to understand that, you know, the system of free speech was developed before that, and so not to focus solely on thinking about free speech only through the lens of the digital age. What came before it is really important to give you some perspective.

So that’s one thing, but I also have two kids, aged 13 and 16, so I’ve thought a lot and fought a lot about some of these issues. I understand where some of the age verification concerns come from. I have parental controls on my children's phones and devices, and try to control it as best as possible, because I do think there can be harms if you spend too much time. But on the other hand, I would also say—and this goes back to the harms and benefits—sometimes there's this analogy that people want to make that social media is like tobacco, which I think is such a poor comparison, because, you know, no one in the world would disagree that tobacco is extremely harmful, right? It's cancerous and all kinds of other things. There are no benefits to tobacco, but social media access, I think, is very different. For instance, I moved to the United States with my family three years ago. My children had no problem speaking English, doing well in school because of YouTube. They could speak almost with the accent, they were immersed into cultural idioms, and they could learn stuff. And also in terms of connections, they have friends back home, it would be very difficult for them to stay in touch the same way that they can now and have connections, if it wasn't due to technology. And so I think that social media for minors also has benefits that make it very, very different from the tobacco analogy. 

Plus, I also think, and here I'm pointing my finger at Jonathan Haidt, that some of the evidence that is being pushed for these kinds of bans seem not to reflect scientific consensus, and that there's a lot of subject matter experts who actually think that the case is much more muddled than than the message that he has pushed in his best selling book, but which is now going the rounds. 

But it amazed me to look at. First of all, let me say I've admired Jonathan Haidt for a long time. I loved his previous work, but I just feel like his crusade on social media for minors and age verification is…in a certain sense, he's gone down some of the roads that he warned against in some of his previous books, in terms of motivated reasoning and confirmation bias and so on. But I saw Jonathan Haidt praise the Minister of Digital Affairs for Indonesia for their age verification bill that is supposed to come into effect now. Indonesia is a country that right now, I think, has a bill in place that will give further powers to the government to ban LGBT content, and what’s the justification? Protecting children. It is a country where someone uploaded a Tiktok video where they said an Islamic prayer before eating pork…two years in prison, right? So it's a country that is in the lower half of Freedom House's Freedom on the Net rankings. So it's amazing to me that a good liberal Democrat like Jonathan Haidt would essentially lend his legitimacy to a country like Indonesia when no one, no serious person, can be in doubt that these kinds of laws will be used and abused by a country like Indonesia to crack down on religious and political, sexual minorities and dissent in general.

JY: Absolutely. And that actually fits really well with something that I've been thinking a lot about too. I know you've written a lot about the Brussels effect and I'm trying to look at the ways in which a similar effect—not necessarily coming from Brussels, of course—is shaping internet regulation in different directions, in terms of laws influencing other laws.

Now, in terms of laws influencing other laws, age verification is, I think, one of the big ones. I mean, seeing these laws modeled after things that the UK or Australia or the U.S. has proposed, and then, just being made so much worse, and then sometimes echoing back here as well. And I think Indonesia is such a great example of that.

JM: Yeah. I mean, Australia sort of opened the Pandora’s box, and everyone is rushing in now, and I think the consequences are likely to be grave, and I think it fits into another issue which I think is even more concerning, that is this rehabilitation or of the concept of digital sovereignty. If you went back 10 years ago and talked about digital sovereignty, you would say, “Well, this is something that they do in China or Russia,” but now digital sovereignty is shouted from the rooftops in Brussels and democracies. 

And you know, I could maybe understand, if digital sovereignty meant, yes, we're going to protect our critical infrastructure, or we don't want to be overly reliant on American tech platforms, given the Trump administration's hostility towards Europe. But digital sovereignty now essentially means a concept of sovereignty which asserts that governments and institutions like the European Union have powers to determine what types of information and ideas their citizens should be confronted with. Now look up Article 19 in the Universal Declaration of Human Rights, what does it say? Everyone has the right to free expression, which includes, and I'm paraphrasing here, the right to share and impart ideas across frontiers, regardless of media, right? You know this. So now we're reverting back to an idea of free expression, which says that the government can now control what type of information that…if a foreign government or information that purports to undermine democratic values in a society, then the government has a right to censor it or require that an intermediary take mitigating steps towards it. I mean, I think that is really a recipe for disaster.

JY: I’m so glad you talked about that. I don’t even think everyone talking about digital sovereignty is working with the same definition. 

JM: No no, digital sovereignty can mean a lot of things. But there’s no doubt that it’s now being stretched to also include pure information and ideas rather than critical infrastructure or industrial policy where it may have a more benign role to play.

JY: Absolutely. Well, we’ve covered a lot of territory, so I’m going to ask you my favorite question, the one we ask everyone: Who is your free speech hero?

JM: I think my free speech hero would be Frederick Douglass. To me, he’s just someone who epitomizes not only being a principled defender of free speech, but someone who did free speech in practice. In his autobiography—he wrote three, I think—but in one of them there’s a foreword by the great abolitionist William Lloyd Garrison, and he describes watching and listening to Frederick Douglass give one of his first public speeches in Nantucket in 1841 and Garrison describes the impact that Douglass had on this crowd and he says something along the lines of: “I think I never hated slavery so much as in that very moment.” So you can almost feel the impact of Douglass’s speech, and that’s the gold standard, right, for what speech can do and why it should be free.

JY: Such a great answer. Thank you.

JM: Thank you.

This is the first installment of a blog series reflecting on the global digital legacy of the 2011 Arab uprisings.

A new generation of protesters, raised on social media and often fluent in the tools of digital dissent, has taken to the streets in recent months and years. In Bangladesh, Iran, Togo, France, Uganda, Nepal, and more than a dozen other countries, young people have harnessed digital tools to mobilize at scale, shape political narratives, and sustain movements that might once have been easier to ignore or suppress.

The tools at their disposal are vast, allowing them to coordinate quickly and turn local grievances into visible, transnational moments of dissent. But each new tactic is met in turn: governments now implement draconian regulations and deploy sophisticated surveillance systems, content manipulation, and automated censorship to pre-empt, predict, and punish collective action. 

This cycle of digital empowerment and repression is not new. In many ways, its roots can be traced to the 2011 uprisings that rippled across the Middle East and North Africa. Often referred to as the “Arab Spring,” these movements didn’t just reshape politics…they transformed how we talk about the internet, and how governments respond in times of protest, crisis, and conflict. Fifteen years later, the legacy of that moment still defines the terms of resistance and control in the digital age.

At the time, we were sold the comforting narrative that the internet would help bring about democracy, that connectivity itself was revolutionary, and that Silicon Valley’s products—particularly social media platforms—were aligned with the people. It was a narrative that tech executives were sometimes happy to amplify and certain Western governments were happy to believe. 

But the same networks that helped protesters to organize and broadcast their demands beyond their own borders laid the groundwork for new forms of repression. Over the years, the same tools that were once celebrated as tools of dissent have become instruments for tracking, harassing, and prosecuting dissenters.

This series examines the digital legacy of the 2011 uprisings that shook the region: how governments refined censorship and surveillance after 2011, how platforms alternately resisted and enabled those efforts, and how a new generation of civil society has pushed back.

"Over the years, the same tools that were once celebrated as tools of dissent have become instruments for tracking, harassing, and prosecuting dissenters."

When Tunisian fruit vendor Mohamed Bouazizi set himself on fire on December 17, 2010, after repeated harassment by local officials, he could not have known the chain reaction his act would spark. After nearly twenty-three years in power, President Zine El Abidine Ben Ali faced a public fed up with repression. Protests spread across Tunisia, ultimately forcing him to flee.

In his final speech, Ben Ali promised reforms: a freer press and fewer internet restrictions. He left before either materialized. For Tunisians, who had lived for years under normalized censorship both online and off, the promises rang hollow.

At the time, Tunisia’s internet controls were among the most restrictive in the world. Reporting by the exiled outlet Nawaat documented a sophisticated filtering regime: DNS tampering, URL blocking, IP filtering, keyword censorship. Yet despite that machinery, Tunisians built a resilient blogging culture, often relying on circumvention tools to push information beyond their borders. When protests began—and before international media caught up—they were ready.

Eleven days after Ben Ali fled, Egyptians took to the streets. International headlines rushed to label it a “Twitter revolution,” mistaking a tool for a movement. Egypt’s government drew a similar conclusion. On January 26, authorities blocked Twitter and Facebook. The next day, they shut down the internet almost entirely, a foreshadowing of what we’d see fifteen years later in Iran.

As Egyptians fought to free their country from President Hosni Mubarak’s autocratic rule, protests swept across the region to Bahrain, where demonstrators gathered at the Pearl Roundabout before facing a brutal crackdown; to Syria, where early calls for reform spiraled into one of the most devastating conflicts of the century; to Morocco, where the February 20 Movement pushed for constitutional change. Outside of the region, movements took shape in Spain, Greece, Portugal, Iceland, the United States, and beyond.

In each context, digital platforms helped circulate images, testimonies, and tactics across borders. They created visibility—and, in turn, inspired a playbook. Governments watched not only their own populations but one another, quickly learning how to disrupt networks, identify organizers, and seize back control of the narrative.

Cause and Effect

To be clear, the internet didn’t create these movements. Decades of repression, corruption, labor organizing, and grassroots activism did. Later research confirmed what many in the region already understood: digital tools helped people share information and coordinate action, but they were neither the spark nor the engine of revolt.

But regardless, the myth of the “Twitter revolution” had consequences. The breathless coverage, and rapid policy reactions that followed shaped state strategy around the world. Governments across the region and well beyond invested heavily in surveillance technologies, developed new legal mechanisms, increased their own social media presence, and found ways to influence platforms. Internet blackouts, once rare, became a normalized tool of crisis response. And companies were forced into increasingly public decisions about whether to resist state pressure or comply.

When it comes to the internet, the legacy of the 2011 uprisings that swept the region and beyond is a story about power: how states moved to consolidate control online, how platforms—often under pressure—have narrowed the space for dissent, and how civil society has been forced to evolve to defend it.

This five-part series will take a deeper look at how the internet as a space for dissent and for hope has changed over the past fifteen years throughout the region and well beyond.  

Sen. Ron Wyden is warning us of an abuse of Section 702:

Wyden took to the Senate floor to deliver a lengthy speech, ostensibly about the since approved (with support of many Democrats) nomination of Joshua Rudd to lead the NSA. Wyden was protesting that nomination, but in the context of Rudd being unwilling to agree to basic constitutional limitations on NSA surveillance. But that’s just a jumping off point ahead of Section 702’s upcoming reauthorization deadline. Buried in the speech is a passage that should set off every alarm bell:

There’s another example of secret law related to Section 702, one that directly affects the privacy rights of Americans. For years, I have asked various administrations to declassify this matter. Thus far they have all refused, although I am still waiting for a response from DNI Gabbard. I strongly believe that this matter can and should be declassified and that Congress needs to debate it openly before Section 702 is reauthorized. In fact, ...

The court found three legal challenges from Maryland cities and one county against oil companies are "far afield" from matters of local or state law.

Climate expert, speaker and author Kate Marvel cited "uncertainty" about NASA climate research as President Donald Trump cuts programs and cancels a lease.

State officials have pledged expedited permitting for the project, which has come under fire from environmentalists.

State legislation would assess insurance policies covering fossil fuel infrastructure and fund projects to improve climate resilience.

Gov. Josh Green called it the state's most serious flooding since 2004.

The number of people displaced from their homes in flooding that started earlier in March has now reached over 34,000, the Interior Ministry said.

Dan Jørgensen said Europe should never again buy "one molecule" of Russian gas.

More coal use now will slow and possibly undermine long-term efforts to phase out coal-fired power in the region.