Insurance claims increased by $10 billion since 1980 due to the disappearance of those ecosystems.

A 9th U.S. Circuit Court of Appeals panel says the link between the orders to unleash fossil fuel and climate injuries is too speculative.

The label would benchmark the environmental performance of data centers, using metrics including energy and water efficiency, renewable energy use, and waste heat recovery.

Scientists said that the goal puts the U.K. on course to meet its 2050 net-zero target, though Tuesday's announcement doesn't include details of how it will be achieved.

JetBio will rely on a diverse range of suppliers, including makers of sugar, second-crop corn as well as waste-based ethanol, for processing into sustainable aviation fuel.

And when the disaster is over, health experts say, try to restore a sense of normalcy by seeking out support, returning to routines and helping others.

Nature Climate Change, Published online: 03 June 2026; doi:10.1038/s41558-026-02660-7

How hailstorms change with warming is not well understood. Here the authors use global projections with different hail proxies to show that hail-prone conditions shift polewards under warming, also shifting crop risk related to hail hazards.

Nature Climate Change, Published online: 03 June 2026; doi:10.1038/s41558-026-02663-4

The authors assess the impacts of China’s Grassland Ecological Compensation Policy on climate and maize yields. They demonstrate reduced temperature and increased precipitation, which are linked to increased crop yields that partially offset restoration costs.

Nature Climate Change, Published online: 03 June 2026; doi:10.1038/s41558-026-02652-7

The Southern Ocean is important for anthropogenic heat uptake, and this regional analysis shows an area with enhanced warming in the high-latitude Indian sector. Model analyses indicate that mesoscale eddies drive upward heat transport, linked to a strengthening of the Antarctic Circumpolar Current.

To accelerate and refine decision-making in a fast-paced, global marketplace, enterprises may deploy generative artificial intelligence models to help summarize and interpret the charts that often fill market summaries and financial reports.

But even the latest vision-language models sometimes struggle with this task, since it requires a model to integrate visual, numerical, and linguistic understanding. A company that invests in a state-of-the-art model might still receive inaccurate or incomplete information.

To fill this performance gap, researchers from MIT and the MIT-IBM Computing Research Lab developed a multifaceted resource for AI users that is specifically designed to teach vision-language models (VLMs) how to effectively interpret charts. 

They used a novel data generation method to build a state-of-the-art dataset that includes more than a million varied charts. The dataset also encodes many visual, linguistic, and numerical components of each chart image, which enable models to robustly reason about the information in a chart.

The researchers used this dataset, called ChartNet, to train a series of open-source VLMs.  Many of these smaller models significantly outperformed orders of magnitude larger, commercial models on tasks like data extraction and chart summarization.

By enabling open-source models to outperform their commercial counterparts, ChartNet could allow small firms with limited budgets to more readily utilize AI. The open-source dataset can be used to improve the capabilities of AI models for tasks like business trend analysis and scientific figure interpretation.

“We developed ChartNet to be a one-stop shop for chart understanding, covering basically anything that an AI model and a practitioner who is training that model might need. We hope our work motivates researchers to achieve state-of-the-art performance with smaller models that don’t require infinite amounts of computation,” says Jovana Kondic, an MIT electrical engineering and computer science (EECS) graduate student and lead author of a paper on ChartNet.

She is joined on the paper by many co-authors from MIT, the MIT-IBM Computing Research Lab, and IBM Research, including Pengyuan Li, a research staff member at IBM Research; Dhiraj Joshi, a senior scientist at IBM Research; Isaac Sanchez, a software engineer at IBM Research; Aude Oliva, director of strategic industry engagement at the MIT Schwarzman College of Computing, MIT director of the MIT-IBM Computing Research Lab, and a senior research scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL); and Rogerio Feris, a principal scientist and manager at the MIT-IBM Computing Research Lab. The research will be presented at IEEE Computer Vision and Pattern Recognition Conference.

A dataset bottleneck

Researchers have made great strides developing generative AI models that excel at natural language processing and reasoning about natural images. But less work has focused on interpreting complex multimodal data contained within charts, Kondic says.

Yet for large and small businesses in nearly every industry, chart understanding is a critical task.

“The finance industry thrives on charts. If vision-language models can extract information out of charts, like descriptions of trends, that facilitates a lot of workflows that happen downstream,” Joshi says.

The lack of high-quality training data is a major bottleneck holding back the development of VLMs that can accurately interpret charts. Many datasets contain limited chart images pulled from the internet and often lack the necessary scale and additional information to help a model interpret the underlying data.

“A vision-language model, unlike our brains, may need to see thousands of examples during training to reliably recognize something as a line chart,” Kondic says.

The researchers sought to overcome those shortcomings by generating synthetic data. Synthetic data are artificially generated by algorithms to mimic the statistical properties of actual data. 

The ChartNet dataset holds more a million high-quality chart images, along with the corresponding code used to generate each chart, a textual description, and a table that contains its numerical information. In addition, each datapoint includes question-and-answer pairs to teach the model how to correctly answer questions about the chart image.

“These additional modes of data guide the model to connect and align the different pieces of information that the chart image encodes,” Kondic says.

Data generation

To build ChartNet, the researchers created a two-step, synthetic data generation pipeline.

First, their automated system translates any pre-existing set of chart images into code. Then the system iteratively augments that code to change different aspects of each chart, such as chart type, data values, topic, colors, etc.

“We can start from a single chart that we use as a seed and come up with hundreds of augmentations of it. This is how we were able to build a dataset with more than a million diverse images,” Kondic explains.

They also incorporated an automated quality check process to ensure the synthetic data are high quality. This process verifies that the code is executable and rendered chart images are accurate and clean.

“We don’t want to just be generating diverse samples. We also want the information to be presented in a meaningful way,” she says.

ChartNet also includes a selection of chart datapoints annotated by human experts. This provides access to additional types of charts and supporting data that carry validity guarantees.

A practitioner could use the annotated data to fine-tune an existing VLM, further boosting performance for a specific application, Joshi adds.

The researchers tested ChartNet by training IBM’s Granite Vision series of models as well as several other open-source models of various sizes and evaluating them on various chart interpretation tasks. The dataset improved the accuracy of all models in chart reconstruction, chart data extraction, chart summarization, and chart question answering. 

With ChartNet, small open-source models consistently outperformed much larger  commercial models. 

“A lot of prior training datasets only focused on answering simple questions about a chart. We tried to go beyond that with ChartNet by generating data that support all aspects of robust chart understanding,” Kondic says.

In the future, the researchers plan to continue expanding ChartNet by incorporating data with added levels of complexity. They also want to draw on feedback from the research community. 

This research was funded, in part, by the MIT-IBM Computing Research Lab.

When a team of MIT students turned up at a national robotics tournament, their robot — aptly named Timbot — wouldn’t work. They’d been invited to demonstrate Timbot at the inaugural United States Governors Cup in Washington, D.C., a March Madness-like competition for high school robotics teams from all 50 states.

Troubleshooting on the fly is par for the course at robotics tournaments. Timbot had a few technical issues, mostly with Wi-Fi, so the team sat cross-legged on the floor and set to work. Meanwhile, high school students started gathering around and asking questions about wiring and subsystems. After about an hour, Timbot was up and running again, scooping up and throwing foam balls as it was designed to do.

“It actually turned into a great moment,” says first-year student Lily Sand. “We ended up tethering the robot with a long Ethernet cable, instead of using wireless, and a lot of students were like, ‘whoa, we do that too!’ It was a nice connection point.”

Leveraging a cultural touchstone for good

Connecting younger students to robotics is one of the MIT students’ goals as members of a new club, FIRSTxMIT, which launched at the beginning of the academic year. Members are all alumni of programs offered by FIRST Robotics (FIRST), a nonprofit that aims to inspire interest in STEM for K-12 students worldwide through team-based robotics programs and competitions.

FIRST has deep roots at MIT. Inventor Dean Kamen collaborated with the late MIT Professor Woodie Flowers, a pioneer in hands-on engineering design education, to establish the FIRST Robotics Competition in 1992. The competition was modeled after the novel robotics competition Flowers had developed for his iconic mechanical engineering class 2.70 (Introduction to Design), which is now 2.007 (Design and Manufacturing I).

Through FIRST, students learn about more than designing, building, and programming robots. The program emphasizes the ethos of “gracious professionalism,” a term coined by Flowers for high-quality work, respect, and cooperation, even in the context of competition. Students also build self-confidence, gain leadership experience, and hone communication skills, as well as technical expertise. 

Many FIRST alumni feel deep gratitude for the program and a strong desire to stay involved. Debbie Ang, co-founder of FIRSTxMIT, still mentors her high school’s team in New Hampshire. Yet, there are few FIRST alumni clubs at universities. Ang and co-founder Perry Han, also a sophomore, met in high school through FIRST and reconnected at MIT. “We noticed that FIRST was founded here, and yet there wasn’t anything organized on campus, even though we kept running into people who had done FIRST and still cared about the community,” she explains.

In fact, participation in FIRST is somewhat of a cultural touchstone among MIT students. MIT associate director of admissions Trinidad Carney, a liaison to FIRST Robotics, estimates that 15-20 percent of undergraduates have participated in the program.

Han and Ang collaborated with Carney to launch FIRSTxMIT, under the auspices of the Edgerton Center, to foster connections among the MIT FIRST community and provide a way for members to channel their passion for FIRST into outreach and public service. Their hunch about the untapped potential an alumni club was spot-on: the kickoff event drew 185 students, and there are about 200 on their Discord channel.

Sharing the “power of FIRST”

Now the club is off and running. They have hosted a gathering for New England FIRST alumni; collaborated with the Josiah Quincy Elementary School in Boston to launch a LEGO Robotics league; volunteered as judges at local competitions; and helped the MIT Admissions Office with outreach. Carney, who advises the club, says, “We’ve actually had other universities reach out to us to say, ‘How did MIT manage to launch a club that’s so successful and compelling?’”

One of the club’s most ambitious undertakings to date was building Timbot, in three days, during Independent Activities Period in January. Robot in 3 Days (Ri3D) is a collegiate challenge in which students build a FIRST Robotics Competition-level robot in 72 hours, a feat that would take about six weeks for a high school team. Experiential Robotics, a consortium that leverages an experiential robotics platform to promote engineering and public service, provided support for MIT’s Ri3D challenge and invited the team to act as STEM ambassadors at the Governors Cup.

In addition to the robotics competition, the two-day event brought together governors and leaders from government, education, industry, and others to underscore the crucial role that states play in supporting STEM education.

To that end, the FIRSTxMIT team demonstrated Timbot, chatted with high schoolers, staffed the MIT Admissions booth, and mingled with VIPs, sharing the value of project-based STEM enrichment opportunities like FIRST. “Having MIT students tell the story of the power of FIRST is incredibly compelling,” says Carney. “They can say: I did this in high school, it shaped who I am, and now I’m at MIT continuing to build and give back.”

A number of governors stopped by the MIT Admissions booth to chat with the students, including Massachusetts Governor Maura Healey. “She talked about the importance of K-12 STEM education and was very supportive,” says Sand, FIRSTxMIT’s logistics coordinator. 

In addition to inspiring others, the MIT students drew inspiration themselves at the Governors Cup. Han recalls speaking to a state senator from Ohio, a former teacher and strong advocate for programs like FIRST. “It really showed me that, when you have people in government that are excited about STEM education, it can really go places.”

Building a better future

Looking forward, Han and Ang plan to take some time to further refine the club’s organization and future goals. Hands-on outreach figures prominently in their plans. “FIRST places a big focus on starting new teams, supporting underserved communities, and spreading awareness,” says Ang. “A lot of us feel that FIRST played a major role in shaping our academic and career paths, so we want to give that opportunity to others.”

“Part of our goal is, we want to put a robot in as many students’ hands as possible to kind of give them a sense that, STEM isn’t just reading the AP Physics C-Mechanics textbook,” Han adds. “It’s actually putting these ideas into practice and building something useful.”

They have no shortage of new ideas they are kicking around, as well. Han is particularly interested in advocating for students to earn Undergraduate Research Opportunities Program or class credit for projects like Ri3D, or for those in the Gordon Engineering Leadership Program to get leadership credit by mentoring a robotics team. He also wants to explore how to leverage FIRST alumni networks to help students with professional development.

Whatever path they take, Carney has no doubt they make an impact. She saw their potential on full display when they built Timbot.

“These students, many of whom hadn’t met before, came from all kinds of backgrounds: different schools, different regions, different life experiences,” she says. “But they worked together with respect, curiosity, and generosity. They’re collaborative, mission-driven, and passionate about making opportunities for others. They make MIT better, and they will make the future better.”

EFF is on the front lines of the fight against tech-enabled tyranny, but we aren't alone. Our team depends on your help to fight back against the surveillance state.

JOIN EFF

People around the world are pushing back against the mass surveillance that undermines privacy and free expression for everyone. You can help during EFF's spring membership drive.

One of the people who joined the fight for digital rights is EFF client Will Freeman. Will created the website DeFlock.me to reveal the dangers of automated license plate readers (ALPRs)—cameras that collect location data on every vehicle they see and upload that to a massive nationwide police database. Deflock.me turns the tables by enlisting ordinary people to track the locations of tens of thousands of ALPR cameras.

But when the police spy-tech company Flock Safety went after Will's website with legal threats citing trademark law, he saw it for what it was: an attempt to silence critics and dim the light on mass surveillance.

The company will try everything it can to downplay the criticism, but EFF will be right there demanding accountability.

"I was totally unprepared to receive a cease & desist letter. I can see how most people would be bullied into submission by a threat like that. That's when I remembered Dave Maass from the EFF introduced himself via email several weeks before, so I reached out for help," Freeman says.

And that's when EFF stepped in. Recognizing DeFlock.me as a quintessential expression of grassroots advocacy and a form of criticism protected by the U.S. First Amendment, EFF's lawyers helped Will fight back. And the Big Surveillance Tech flinched.

But these battles against Flock's Spying tools rage on. In cities around the country, privacy advocates are pressuring officials to block or end contracts for ALPRs—and winning. The company will try everything it can to downplay the criticism, but EFF will be right there demanding accountability.

Get the new Claw Back member t-shirt featuring a fierce feline swatting at community surveillance. You might empathize with him, but there’s a better way. Let’s end the law enforcement contracts, harmful practices, and twisted logic that enable mass spying in the first place.

"I'm really grateful the EFF was able to step in and help. Without them, free speech would be only for those wealthy enough to defend themselves against billion dollar companies. We've grown a lot since then and are expanding our efforts to expose and push back against mass surveillance on our streets," Freeman says.

Support the movement

stop mass surveillance tech today when you join EFF

____________________

EFF is a member-supported U.S. 501(c)(3) organization. We've received top ratings from the nonprofit watchdog Charity Navigator since 2013! Your donation is tax-deductible as allowed by law.

As part of their 20th Anniversary celebration, Dark Reading asked five cybersecurity industry leaders who wrote blogs or columns for them over the years to select their favorite piece and share their reflections on the topic today. This is my section.

Renowned technologist and author Bruce Schneier contributed a column on June 20, 2010, warning about cryptography’s inability to secure modern networks, a point he says he has been trying to argue since 2000.

“For a while now, I’ve pointed out that cryptography is singularly ill-suited to solve the major network security problems of today: denial-of-service attacks, website defacement, theft of credit card numbers, identity theft, viruses and worms, DNS attacks, network penetration, and so on...

An anonymous security researcher called “Nightmare Eclipse” has been publishing a series of significant security exploits against Microsoft Windows—including one that breaks BitLocker. Microsoft has threatened legal action against the researcher. Lots of recriminations are being traded back and forth.

Emergency officials are searching for loopholes to save lives as President Donald Trump slashes programs.

The ruling Monday is a blow to the Trump administration's efforts to dismantle the National Center for Atmospheric Research.

The 1,250-megawatt line will deliver much needed renewable energy to New York City.

The case is contractual in nature, meaning the states can only seek damages in another court, ruled the judge.

Energy-hungry data centers must support the EU’s shift to carbon-free power, says Energy Commissioner Dan Jørgensen.

Gigascale is backing a number of clean tech startups supporting the artificial intelligence boom. It has previously invested in technologies ranging from wave energy to laser fusion systems.