Computer science student Jack Zemke was getting nowhere despite sending out hundreds of job applications that mostly disappeared into the void—ignored.

With his graduation from Tulane University just days away, he was considering moving back in with his parents when he got a break thanks to a group project he had worked on involving an AI-powered search tool. As he touted it at an engineering conference, an employee at a civil engineering firm approached.

“We’re not like the most on top of trends, but we know that AI is becoming a big thing,” he recalled the man saying. “We need somebody to be essentially an evangelist.”

Zemke responded: “I’m your man.”

It has been a difficult job market for new graduates, especially in fields like software development where artificial intelligence is shouldering more of the work. While college students majoring in computer science during the recent coding boom were getting job offers without even applying, graduates today in philosophy, art history and ethnic studies majors all have higher employment rates.

But as the job market evolves with AI and entry-level jobs become harder to find, students and schools are adapting. Colleges are experimenting with business-specific internships and bolstering training in AI tools. Schools say most graduates ultimately are getting jobs, although increasingly, the employers hiring aren’t tech giants but smaller businesses looking for help figuring out what to do with AI.

Such was the case with Zemke, 23, who spent a year at the engineering firm before getting hired this summer at an AI tech startup.

“These jobs are expecting you to be able to jump in and kind of operate on the frontier,” Zemke said. “There’s so much churn. People are getting laid off. People are leaving. It really feels like a total frenzy right now.”

A coding boom that enticed many students has fizzled

The elevated jobless rate in the field—at around 7.1% for recent computer science and computer engineering graduates, according to an Associated Press analysis—reflects in part the huge numbers of students who enrolled during the coding boom.

Hiring surged during the pandemic as everything went digital, with software development postings on the job search site Indeed peaking in early 2022.

But now tech giants like Microsoft, Meta and Google are turning to AI for a larger share of the code writing, and big employers like Amazon are shedding workers as spending on AI accelerates. Far fewer jobs in software development are being posted, according to a Federal Reserve analysis of the Indeed data.

For workers in their early 20s in what are being called “AI-exposed” occupations—such as software development—employment is 19% below where it would be if it had kept pace with employment in fields less exposed to being reshaped by AI, according to the Stanford Digital Economy Lab. The U.S. Census Bureau likened the employment decline to finishing college during a large recession.

What the industry is demanding now is workers who can oversee AI tools and not blindly accept their outputs—skills more common in seasoned workers, said Tracy Camp, the Computing Research Association’s executive director and CEO.

The effect on hiring of young workers is so concerning that the association has reached out to industry leaders, urging them to invest in entry-level workers.

“If they don’t change how hiring is currently happening, they’re not going to have mid-level career people in a few years,” Camp said.

Students also are taking note. After a decade that saw the number of recent computer science and computer engineering graduates triple, to about 362,000 in 2024, enrollment began to drop last year. The latest data from the National Student Clearinghouse Research Center shows that by spring this year, computer and information science enrollment has fallen by 8.4% at four-year institutions.

A shift in hiring disadvantages international students

For international students, who have enrolled in high numbers in computer science programs, it has been an especially challenging job market. Their struggles owe partly to changes to U.S. immigration policy as well as the shift in the kinds of available jobs.

Samir Khuller, the chair of Northwestern University’s computer science department, said many of the smaller companies that are hiring don’t have big human resources departments with experience handling H-1B work visas.

International students generally need an H-1B work visa to stay in the U.S., which requires sponsorship from an employer. The program has become a target of the Trump administration, which is seeking to raise the fee for the visas to over $100,000.

Computer science graduate Collins Kibet, from Kenya, has sent out hundreds of applications and landed just a handful of interviews—mostly with AI bots. The Southern New Hampshire University graduate said that during the rare job interview with a person, the tone shifts when employers learn they would have to sponsor his H-1B work visa.

“Everybody just keeps quiet,” said Kibet, who started his own software company because he could not find a job.

Colleges are rethinking training for computer science students

Watching students navigate job searches has been hard for Nicholas Mattei, an associate computer science professor at Tulane. When he looks at some of the job ads, he thinks, “Really?”

“All the companies are coming back to us and saying, ‘Oh, well you need to train these people even better, and they need to be able to jump into a middle management role as opposed to an introductory or junior developer role,'” said Mattei, who also chairs a committee studying AI for the Association for Computing Machinery. “And it’s kind of crazy.”

Many colleges are retooling their computer science programs. With entry-level jobs going away, they are tasked with preparing students to do more.

To give students a leg up, professors at Georgia Tech researched what AT&T wanted and trained students for a month before they started internships with the company. The school plans to expand the effort for a program it will call “Bootcamp to Industry,” said Fisayo Omojokun, the associate dean for undergraduate education in the school’s College of Computing.

Staff tasked with recruiting businesses for College of Computing job fairs are also reaching out to a broader range of industries.

“Earlier this year, we got a call from a forestry company of all things,'” Omojokun said. “They’re like, ‘Hey, we are in forestry, and we know that AI can help in what we do in different ways.'”

Who’s behind this story?

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

© 2026 The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed without permission.

As Eastern Europe’s power sector has taken a hit from a record drought, Bulgaria has benefited from building battery storage capacity alongside its rapid solar expansion.

Around 100 battery energy storage system (BESS) sites are now operating, most of them built in the past two years.

The large-scale projects take advantage of former industrial sites and airfields or are integrated into existing power stations, including coal-fired plants.

“The scale and speed of storage deployment in Bulgaria is unprecedented in a European context,” Stavros Papathanassiou, a professor at the National Technical University of Athens, told AFP.

Low administrative hurdles, a receptive market for clean energy and EU funding are among the reasons experts cite for the country’s leading role in the region.

The battery installations absorb surplus power from solar installations and feed it back into the grid when demand rises.

That proved a boon when exceptionally low water levels in the Danube disrupted power generation at nuclear and hydroelectric plants in Eastern Europe, boosting demand for electricity imports.

In July and August, Bulgaria—one of the EU’s poorest countries—recorded net electricity exports of around 830 gigawatt-hours a month, more than twice the average in May and June.

In July alone, Romania accounted for nearly half of Bulgaria’s outward electricity transfers, valued at an estimated 74 million euros ($85 million).

Reducing coal reliance

At Maritsa East 3, once Bulgaria’s second-largest coal-fired power plant, the shift away from dependence on the fuel is visible on the ground.

Built under communism with four generating units, the plant sharply reduced coal operations in 2024, repurposing existing grid infrastructure for a battery facility.

“We could leverage the large connection to the grid. We decided to disconnect one coal unit and connect batteries,” said Vassil Shtonov, head of Bulgaria operations for the British-based power company ContourGlobal.

“Coal is out, battery is in—it was the right place at the right time,” Shtonov told AFP, though he said Bulgaria should maintain a minimum coal-fired reserve for severe winters or disruptions to gas supplies.

While Bulgaria accounted for about 9% of European battery-storage additions and about 2% of solar additions in 2025, it can “serve as a positive example for other European nations,” said Lennart Herrmann at the Munich-based research and consulting firm FfE.

“The right incentives … can enable a swift energy transition,” he told AFP.

The Bulgarian energy ministry said the country had around 5.7 gigawatts of operational battery-storage power capacity this month.

That is equivalent, in power terms, to nearly a quarter of Bulgaria’s installed power-generation capacity overall—and more than five times neighboring Romania’s operational battery capacity.

Businesses in Bulgaria have also signed contracts for nearly 14 gigawatt-hours of large-scale storage, part of which has already come online.

“The aim is for storage systems to find a sustainable role in the market,” the ministry said, adding that it “must go hand in hand with other investments, such as in transmission and distribution grids.”

‘By chance’

Nikola Gazdov, president of Bulgaria’s Association for Production, Storage and Trading of Electricity, attributed the battery boom to a combination of factors rather than a crafted strategy.

“Everything happened almost by chance,” he told AFP.

Solar and wind capacity has expanded rapidly across the region, but Bulgaria’s renewables boom has been particularly solar-heavy, creating growing daytime surpluses.

And in Bulgaria, large battery projects can be up and running in just nine to 18 months, much less than many solar or wind projects, Gazdov added.

But battery storage technology has its limits.

One is that grid connections with neighboring countries are limited in the amount of current they can transport.

Another is that current batteries can store power only for relatively short periods, “typically between one and six, sometimes up to eight hours,” said Herrmann at FfE.

So while they can support the grid in case of a temporary drop in power supply, they cannot replace a nuclear reactor that goes offline due to drought.

And their success could ultimately reduce their own profitability, because when more BESS operators buy electricity when prices are low and sell it back at peak times, the gap between the two price periods will narrow.

Who’s behind this story?

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

© 2026 AFP

A rogue OpenAI model bypassed safeguards during training and hacked an Australian government website, Prime Minister Anthony Albanese said as he admonished the ChatGPT creator over an “obviously unacceptable” breach.

The AI tool sought access to a health statistics portal in June and “didn’t accept no for an answer,” sidestepping restrictions to breach a section hosting private files, Albanese said.

OpenAI did not raise the alarm with the Australian government until September, when it sent a message to a generic email inbox that is checked only once a day.

Global worries about the power of advanced AI tools are mounting after a string of hacking incidents involving models from both OpenAI and rival developer Anthropic.

“Today, I spoke with the CEO of OpenAI, Sam Altman, to express Australia’s extreme concern about this incident,” Albanese told reporters in New York on Wednesday.

“I also expressed my disappointment that it took the company way too long to inform the government what had occurred,” he said.

“It took until September 10 before there was any notification at all—and the notification was an email sent to just the public mailbox.”

The AI tool accessed public and nonpublic files hosted on an old health statistics website.

Albanese said there was “no evidence” that personal information had been accessed and that other government services had been compromised.

“Nonetheless, this situation is obviously unacceptable.”

The breach occurred in June as leading artificial intelligence company OpenAI ran training exercises to rate the performance of AI models.

It asked the model to search the internet for data showing how much the Australian government spent on medicine, Government Services Minister Katy Gallagher told reporters.

OpenAI said it did not spot the rogue activity until August, when it reviewed what the AI tool had been doing.

Rogue AI fears

The San Francisco-based company did not alert the Australian government until Sept. 10, when it sent an email to a generic government inbox.

“That email address is looked at once a day,” Gallagher said. “We have someone who goes and has a look through. It sometimes gets a number of notifications, sometimes many of them are hoaxes.”

Defense Minister Richard Marles said the AI model had “scaled the fence.”

“It asked a question, the information was not given and rather than leaving at that point, it scaled the fence.”

Australia has launched a rapid review of the incident, which will include the national intelligence agency responsible for cybersecurity.

OpenAI said it spotted the breach while carrying out an “extensive review” of its AI models.

“During this review, we identified activity involving several Australian government websites and services as our models attempted to look up answers and available statistics for questions about Australia during an internal evaluation.

“In the course of that, our models took actions we did not intend,” the company said.

Altman and other tech CEOs addressed a special meeting of the U.N. Security Council on Wednesday about AI risks.

More than 100 organizations around the world, including OpenAI and Anthropic, signed an open letter last month calling for a global effort to “strengthen cyber defenses” against AI-powered cybersecurity threats.

It came after two OpenAI models escaped from a closed testing environment and broke into the internal systems of Hugging Face, a site that AI developers use to store and share code.

Anthropic recently discovered that its models had gained unauthorized access to three unidentified organizations during testing that was supposed to keep them away from “real-world” systems.

Google’s consumer AI model Gemini hacked multiple systems by guessing login credentials, the company said last week.

Who’s behind this story?

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

© 2026 AFP

Leaders of major AI firms told the U.N. Security Council on Wednesday that the technology calls for heightened vigilance, with Anthropic’s CEO pledging to slow down where needed to ensure the safety of new releases.

“This moment calls for extreme care,” OpenAI CEO Sam Altman said in New York at an emergency meeting of the 15-member council.

He warned that as AI systems become more capable and autonomous, they could move faster than institutions or concentrate power in too few hands.

Altman was among industry leaders briefing the council following weeks of global headlines about the risks of uncontrollable artificial intelligence and calls for a slowdown, which have been resisted by U.S. President Donald Trump.

“The industry must not accept too much technological risk just because the benefits are too great and important to slow down,” Altman told the United Nations’ top body.

He also urged governments to work with each other and the industry on decisions shaping the technology’s future.

“In our history, there have been times when countries that compete and don’t always like each other very much still come together for shared interests and the collective good in the face of a powerful new technology,” he said.

“We believe this must be one of those times,” he added.

Slowing down

Speaking after Altman, Anthropic CEO Dario Amodei said his company understands that as AI models become more powerful, “even more stringent standards of safety will be required.”

He said Anthropic aims to improve its ability to prevent misuse of the technology.

“We will slow down as much as necessary in order to make sure that every successive AI technology that we release is actually safe,” Amodei said by video.

“If managed poorly, I even believe that AI could be a risk to humanity as a whole,” Amodei added.

Other speakers included Clement Delangue of Hugging Face, which was the target of a recent cyberattack by autonomous OpenAI software.

This marked the first in a series of incidents that have raised alarm over the rapidly evolving technology.

In early September, Jacob Coxon, a researcher who worked at both OpenAI and Anthropic, decided to leave the industry and accused both companies of “gambling with our lives” in the race to develop AI models capable of self-improvement.

‘Middle path’

The executives’ remarks come shortly before Trump is due to welcome Chinese leader Xi Jinping for a summit where AI will be on the agenda.

On Wednesday, Altman said his company would try to walk a “middle path.”

“We don’t want to fall into the trap of blind optimism,” he said, but he also cautioned against falling into “doomerism.”

Yoshua Bengio, the Canadian computer scientist considered one of the founders of AI, told the Security Council that “more and more people are rightly worried about recent advances.”

“We must channel that concern into productive action,” he said.

Delangue of Hugging Face noted that while his company was targeted by AI, it also used the technology to fend off the attacks. “The world needs open-source AI more than ever to defend itself.”

Altman, Amodei and several other leading AI executives have been calling for global oversight of the technology.

OpenAI has urged the United States to lead a global effort to establish standards for AI, noting that its industry is at the technical frontier.

Prime Minister Andy Burnham said Tuesday that Britain would launch an effort to set global AI standards during its 2027 G20 presidency.

French Foreign Minister Jean-Noel Barrot similarly urged joint action within the U.N. framework and with companies to set rules for a new era.

But diplomatic efforts could clash with the U.S. position.

Michael Kratsios, director of the White House Office of Science and Technology Policy, said rapid advancement is not a reason to pause AI’s development or constrain it with new governance structures.

Trump on Tuesday rejected international regulation of AI, calling such proposals a “globalist scheme.”

Who’s behind this story?

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

© 2026 AFP

Imagine if your car weren’t powered by a conventional battery under the hood but by energy stored in its walls or even its roof? This is a future that Dr. Bhagya Dharmasiri and her colleagues at the Deakin Institute for Frontier Materials have been exploring, bringing load-bearing structures and rechargeable energy storage together to create structural battery composites.

They’re working to embed zinc-based energy storage into carbon fiber structures rather than use the lithium-ion batteries found in many modern electronics. The goal is to make safer, better-performing electric cars, airplanes and more.

Rethinking strategic metals

Lithium and zinc are both critical metals for energy storage, but they have different advantages and risks. Although lithium-ion batteries can deliver a large amount of energy, they also pose a risk of fire or toxic gas release if handled incorrectly. These risks can lead to explosions and fires, as seen in recent years with portable chargers and e-bikes.

There are several approaches to mitigating these risks, including safer electrolytes and improved materials and battery design. Dharmasiri and her colleagues are also researching these areas.

“Lithium remains an important energy-storage technology, and there are many ways we can make these systems safer. Zinc offers another promising approach, particularly where safety and sustainability are priorities,” Dharmasiri said.

While zinc stores less energy than lithium, it can use water-based electrolytes, reducing reliance on flammable components. Zinc is also abundant and recyclable, making it an attractive complementary option for safer, more sustainable structural batteries.

Embedding energy storage into carbon fiber

Renowned for its lightweight strength, carbon fiber is an ideal building material for structural battery composites, particularly in weight-sensitive sectors like electric vehicles, aerospace and defense.

As Dharmasiri explains, “Traditionally, carbon fiber composites are designed to do one main job: carry mechanical loads while keeping structures lightweight. We are asking whether that same material can do more.”

In a paper recently published in the journal Composites Part B: Engineering, Dharmasiri and her Deakin colleagues Dr. James Randall and Professor Luke Henderson presented a scalable approach to redesigning conventional carbon fiber composites into multifunctional zinc structural batteries.

They do this by modifying carbon fiber surfaces to host electrochemically active materials for energy storage.

“Zinc is particularly interesting for structural batteries because performance is not simply about achieving the highest possible battery energy density,” Dharmasiri says. “A structural battery performs two functions simultaneously, so we need to consider what it contributes to the performance, safety and efficiency of the whole structure.”

At its core, this redesign process involves electrochemical surface modification of the carbon fiber, integrating electroactive zinc and manganese dioxide while maintaining mechanical integrity.

Materials of the future

Dharmasiri hopes to move from laboratory-scale zinc structural battery composites to building engineering components in which structural performance, energy storage, safety and multifunctionality are optimized together.

Reducing the number and weight of separate components could create lighter cars, airplanes and defense craft with greater range, endurance and functionality.

“The same structure could potentially store energy, sense damage and provide electromagnetic protection. Rather than continuously adding separate components for every required task, we could design the material itself to perform several of those functions,” Dharmasiri said.

By bringing chemistry and materials engineering together, Dharmasiri and her colleagues aim to rethink what a structure can do.

Who’s behind this story?

Lisa Lock

BA art history, MA material culture. Former museum editor, paramedic, and transplant coordinator. Editing for Science X since 2021.

Full profile →

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

Autonomous vehicles and robots need to understand not only what is around them, but how objects and people are moving. A cyclist crossing the road, a car slowing down, a pedestrian stepping off the curb, and a parked vehicle all create different motion cues that a machine must interpret quickly and accurately.

A research team led by assistant professor Zhao Na from the Singapore University of Technology and Design (SUTD) has developed IterFlow, a lightweight learning framework that helps 4D radar estimate the 3D motion of points in a traffic scene. Their study, posted to the arXiv preprint server, addresses a key challenge in autonomous perception: how to make radar-based motion understanding more accurate without depending on expensive LiDAR-based supervision.

Refining motion estimates from sparse radar

4D radar is attracting growing interest because it is more compact, more cost-effective and more robust in adverse environmental conditions than LiDAR, which is short for light detection and ranging. However, radar point clouds are also sparse and noisy, making it difficult for AI systems to estimate scene flow—the 3D motion of points between consecutive sensor frames.

IterFlow was developed to tackle this problem through a more focused design. The research team designed a task-specific network with a concise training strategy to improve radar scene flow performance. Rather than relying on increasingly complex multitask systems, IterFlow refines motion estimates step by step and uses targeted training signals to reduce errors in sparse radar data.

A central feature of IterFlow is that it does not require LiDAR-based pseudo scene flow labels during training. Instead, it uses RGB images and odometry—information about the vehicle’s own movement—as auxiliary supervision. At test time, the system needs only radar point clouds as input.

“IterFlow shows that better radar scene flow estimation does not have to depend on increasingly complex models or costly LiDAR supervision. By using images and odometry during training, we can make radar-based motion understanding lighter, more cost-effective and more applicable to real-world autonomous systems,” Zhao said.

Keeping moving objects distinct

In everyday terms, scene flow estimation helps a system work out how each point captured by a sensor is moving from one moment to the next. In a road scene, this could mean estimating the motion of points belonging to moving cars, cyclists or pedestrians while distinguishing them from static background points such as parked vehicles or roadside structures.

The team’s method uses camera images to provide object-level guidance. Through 2D tracking and segmentation, the system identifies object instances in images and projects this information into 3D radar space. This helps reduce a common source of error in radar scene flow learning, in which a model may mismatch moving foreground points with static background points.

For example, if radar points are sparse, a moving cyclist and nearby static background points may appear close together in the data. Existing methods that rely mainly on spatial distance may wrongly encourage these points to move in similar ways. IterFlow’s instance-aware losses reduce this problem by applying motion consistency within the same object instance, rather than across points that are merely nearby.

IterFlow also uses a ball query-based grouping method that is better suited to sparse radar data. Unlike k-nearest-neighbor methods, which always return a fixed number of neighbors even if some are far away, ball query first checks whether points fall within a defined spatial radius. This helps avoid false correspondences in sparse radar regions and improves robustness.

Efficiency gains and remaining limits

Experiments on the real-world View-of-Delft dataset showed that IterFlow outperformed the previous radar-based cross-modal scene flow method CMFlow while using only three losses, about 40 times fewer parameters and about 30 times fewer giga floating-point operations, or GFLOPs, a measure of computational cost. The results suggest that radar scene flow estimation can be improved without adding costly sensors or substantially increasing model complexity.

For now, this research remains at the experimental stage. The current method uses a PointNet++ point cloud feature extraction network, which supports only a fixed input point cloud size. Future work will focus on overcoming this limitation before the approach can be more broadly tested in practical vehicle or robotic systems.

By reducing reliance on costly LiDAR supervision and improving how radar learns motion from sparse data, IterFlow points toward more efficient radar-based perception for autonomous systems. Its broader significance lies not in replacing other sensors immediately, but in showing how lower-cost sensing can be made more capable through carefully designed machine learning.

Who’s behind this story?

Lisa Lock

BA art history, MA material culture. Former museum editor, paramedic, and transplant coordinator. Editing for Science X since 2021.

Full profile →

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

While lithium-ion batteries (LIBs) remain the most widely used rechargeable batteries worldwide, energy engineers have been testing various alternatives with different underlying chemistries. These include lithium-sulfur batteries, which store and release energy by moving lithium ions between two electrodes on opposite sides of a cell, while sulfur undergoes reactions in the cathode (i.e., positive electrode).

Lithium-sulfur batteries could offer several advantages, including high energy densities and lower production costs, because sulfur is abundant and can store significant charge relative to its mass. Despite their potential, these batteries often exhibit low operating voltages (i.e., the pressure driving current), slow electron-transfer reactions, and energy losses caused by the migration of sulfur compounds between electrodes.

Researchers at the University of Maryland, Vanderbilt University, the Brookhaven National Laboratory, and other institutes recently introduced a new ionic liquid electrolyte that could improve the performance of lithium-sulfur batteries. This electrolyte, presented in a paper published in Nature Energy, was found to increase both the voltage and energy storage of lithium-sulfur batteries.

“Our starting question was simple: could we get more energy from sulfur, an abundant and inexpensive material?” Prof. Chunsheng Wang, senior author of the paper, told Tech Xplore. “Earlier work in my group on chlorine- and bromine-based battery materials inspired us to use halogens to oxidize sulfur at a high potential, further increasing the energy density.”

Overcoming the limits of conventional lithium-sulfur batteries

In conventional lithium-sulfur batteries, every sulfur atom gains or releases two electrons at the cathode during charging and discharging, while these electrons flow through the battery’s external electrical circuit. Prof. Wang and his colleagues wanted to make a third electron available for transfer, thereby allowing the same amount of sulfur to store more charge at a higher voltage.

“Crucially, this extra reaction needed to work repeatedly without continually consuming the liquid inside the battery,” explained Dr. Nan Zhang, the paper’s first author and postdoctoral researcher in Dr. Wang’s group.

As part of their study, the researchers tested various electrolytes with different proportions of lithium salt and an ionic liquid containing chloride. Professor De-en Jiang and his PhD student Jinyi Zhang at Vanderbilt University then ran computer simulations to model the movement and interactions of atoms and molecules in batteries with these electrolytes. Ultimately, the team identified the best-performing electrolyte and used it to design a lithium-sulfur battery with a different internal chemistry.

“We combined lithium metal on one side of the battery with sulfur, porous carbon, and lithium chloride on the other,” said Dr. Jijian Xu. “We designed it so that chloride ions, a charged form of chlorine, could react with sulfur during charging to form disulfur dichloride. This opens up the additional energy-storing reaction.”

Notably, the disulfur dichloride produced while lithium-sulfur batteries are charging does not mix readily with the team’s electrolyte. This helps keep the compound within the porous carbon cathode, preventing it from shuttling across the battery and reacting with lithium metal, which would waste stored energy.

“We tested how much charge the batteries could store, how they performed at different charging and discharging speeds, and how they performed over repeated use,” said Prof. Wang. “We also tested a flat, pouch-shaped prototype. X-ray and laser-based measurements tracked the chemical changes. Computer calculations and simulations, including density functional theory (DFT) and molecular dynamics (MD), helped us understand the reactions and how the electrolyte worked at the atomic level.”

Towards lithium-sulfur batteries with higher voltages and capacities

In initial evaluations, the team’s proposed electrolyte and battery chemistry achieved very promising results. Their design successfully produced a redox reaction involving three instead of two electrons per sulfur atom, boosting the resulting battery’s capacity and average operating voltage.

“Our main achievement is getting more energy from sulfur by using this additional reaction, not simply by putting more sulfur into the battery,” said Prof. Wang. “In our tests, the charge delivered per gram of sulfur increased by about 58%, and the average voltage rose from 2.05 to 2.54 volts, compared with the conventional lithium-sulfur battery used in our study.”

The researchers also estimated that increasing the sulfur cathode’s areal capacity (i.e., the amount of charge stored per unit area) could improve the battery’s stack-level specific energy. They used their approach to create an initial pouch-shaped prototype and assessed its performance in the lab.

“The pouch-shaped prototype retained 78% of its original charge-storage capacity after 100 charge-and-discharge cycles,” said Prof. Wang. “The sulfur-carbon materials delivered more than 1,700 watt-hours per kilogram.”

In the future, the team’s proposed battery chemistry and the electrolyte they introduced could contribute to the realization of more reliable lithium-sulfur batteries with high capacities and higher voltages. Meanwhile, Prof. Wang and his colleagues plan to continue refining their electrolyte’s underlying chemistry to achieve even better results.

“These findings demonstrate that high-valent sulfur redox is viable, opening a new route toward ultra-high-energy-density batteries,” added Prof. Wang. “We now want to improve the electrolyte so the battery lasts longer and works better at faster charging and discharging speeds. We also want to reduce the unwanted reactions that gradually consume this liquid. The next challenge is to turn the chemistry we have demonstrated into a more practical rechargeable battery.”

Who’s behind this story?

Ingrid Fadelli

Freelance journalist with BSc Psychology and MA International Journalism. Covers AI, robotics, neuroscience, and astrophysics since 2018.

Full profile →

Robert Egan

Bachelor’s in mathematical biology, Master’s in creative writing. Well-traveled with unique perspectives on science and language.

Full profile →

© 2026 Science X Network

NASA researchers know that when you settle into your seat on a commercial flight, you expect a smooth takeoff, views over the clouds, a steady descent and, hopefully, an early arrival.

But when your flight is delayed on the tarmac instead of taking off, or it ends up in a holding pattern rather than landing on time, things change. Your experience goes from smooth to anxiety-inducing as you worry about making your connection or getting home in time for dinner.

Large airports are among the busiest, most complex environments in aviation, with aircraft, ground crews and service vehicles sharing crowded taxiways. Researchers at NASA’s Ames Research Center in California’s Silicon Valley recently worked with Boeing to advance three types of field tests—digital taxi information, safe taxiway and safe runways—that could lead to safer, more efficient runway operations at airports.

During the digital taxi tests, pilots received taxiway guidance directly on cockpit displays or tablets instead of verbally from air traffic controllers. Aircraft autonomously followed digital routes while researchers monitored a suite of sensors designed to identify vehicles or other aircraft obstructing the taxi path and runway. The system reduced pilot and air traffic controller workloads and the risk of verbal errors.

Safe runway technology testing can also improve situational awareness for approaching aircraft. During a test involving a Boeing aircraft preparing to land, the same sensors flagged a vehicle on the runway, giving pilots additional awareness to help them avoid potential collisions or other safety concerns.

Together, these NASA capabilities aim to reduce miscommunication, ease pilot workloads and keep airport traffic moving smoothly. Future testing will integrate the sensor and digital taxi systems into a simulated air traffic control environment to evaluate how the technologies can benefit overall management of the airspace.

For years, NASA has worked to improve your flying experience by developing new technologies to modernize the commercial airline system. Key NASA technologies streamline and digitize the flying experience—from the departure gate to the skies to your safe arrival at your destination.

“Aviation safety is key to NASA’s research,” said Parimal Kopardekar, director of NASA’s Airspace Operations and Safety project. “Technology that can provide additional autonomy and support a future airspace with multiple aircraft operating in harmony is key to advancing the National Airspace System.”

NASA’s research innovations continue after your flight takes off. Modern flights constantly respond to shifting weather, turbulence and traffic. Even small changes in direction or altitude can affect when a plane arrives. These changes can force flights into holding patterns while air traffic controllers attempt to rebalance the busy airspace.

NASA’s air traffic management researchers have been working for years to reduce those situations. In a 2025 collaborative effort with Boeing, United Airlines and international partners, NASA evaluated real-time trajectory sharing on domestic and transoceanic flights.

During that testing, a United Airlines Boeing 737 aircraft shared frequent flight information with airline operations centers and air traffic control. NASA used the data to understand how frequently those updates should be sent and what details matter most for generating accurate arrival predictions.

Better information helps controllers sequence traffic more precisely, which means fewer holding patterns and more direct descents for passengers.

Pre-departure rerouting technology and digital exchange tools developed at NASA allow dispatchers and controllers to see the same digital picture of flights preparing to depart.

When a better route becomes available, controllers could coordinate the change digitally instead of relying on verbal communication between pilots, controllers and dispatchers. The technology could lead to fewer delays, reduced fuel consumption and more predictable operations for passengers.

NASA has now transferred the routing technology to the Federal Aviation Administration (FAA), and airlines will continue to test it. These tools build on decades of NASA contributions to national airspace modernization.

In coordination with the FAA, NASA has advanced automation concepts, improved how arrival and departure flows are managed, and introduced data-driven software that commercial airlines use every day.

By working closely with airlines, manufacturers and global partners, NASA is helping to improve every phase of flight to make air travel safer and more reliable, now and in the future.

Who’s behind this story?

Swati Mestri

Swati Mestri holds a bachelor’s degree in Electronics Engineering and has worked as a content editor since 2019. She has experience editing research documents across technology, health care, and materials science, and has a particular interest in technology and space.

Full profile →

Andrew Zinin

Master’s in physics with research experience. Long-time science news enthusiast. Plays key role in Science X’s editorial success.

Full profile →

Artificial intelligence models can give users the wrong answer and do so with great confidence. They can also hedge and warn that they are unsure—even when they get the answer right.

A study led by UC Riverside computer scientists helps explain why. Researchers found that confidence and correctness can arise from different internal features within large language models, challenging the assumption that a model’s confidence reliably indicates whether its answer is accurate.

Their discovery, published on the arXiv preprint server, could help build more reliable AI models. As large language models are increasingly used to inform decisions and complete tasks, developers need better ways to determine when their answers can be trusted.

By identifying internal features associated separately with confidence and correctness, the UCR-led research points toward ways AI systems could be adjusted so they are more confident when they are right and more cautious when they are likely to be wrong.

“The main assumption in the field is that when the model is confident, it is likely to be correct, and when the model is unsure, it is more likely to be incorrect,” said Het Patel, a UCR computer science doctoral student and lead author of the study. “But we often see the counterexamples that are well documented. A model can answer with certainty, but also be wrong, or it can answer while being less confident and can be correct.”

The researchers went beyond documenting that mismatch. They identified internal features associated with confidence and correctness and showed that altering some of them could change model behavior without the costly process of retraining an entire model.

Separating confidence from correctness

Patel explained that modern AI models are created by training enormous networks of mathematical units on vast amounts of data. During training, the network repeatedly adjusts billions of numerical parameters, called weights, as it learns patterns in the data. In a language model, those learned patterns allow it to predict which words are likely to follow others and ultimately generate responses to questions.

Patel and his colleagues wanted to know what was happening inside the models when confidence and correctness did not match.

They studied two “open-weight” large language models—Meta’s Llama-3.1-8B and Google’s Gemma-2-9B—whose internal workings researchers can examine. Using multiple-choice questions, they separated responses into four groups based on whether answers were correct or incorrect and whether the models were confident or uncertain.

They then used tools called “sparse autoencoders” to examine the models’ internal activity. That allowed them to determine which features became active with particular behaviors.

The analysis identified three kinds of features: those associated primarily with uncertainty, those associated primarily with incorrect answers, and “confounded” features associated with both. The researchers then suppressed selected features as the models answered questions. Patel compared the process to “turning knobs” inside a model to see how its behavior changed.

Different features, different effects

The differences were striking. Turning off features associated purely with uncertainty sharply reduced accuracy, suggesting those features play an important role in producing good answers. By contrast, suppressing most features associated solely with incorrect answers had little effect.

The confounded features produced a different result. Suppressing features associated with both uncertainty and incorrectness improved accuracy by up to 1.1% while reducing the models’ uncertainty by up to 75%. Similar effects appeared across different question-answering benchmarks.

The interventions are made when an already-trained model answers questions and do not require retraining it.

“In a sense, it’s kind of like adjusting or modifying the values of these activations or features after the fact to kind of get the behavior you want,” Patel said.

Internal signals flag risky answers

Another experiment suggests the internal signals could eventually help AI systems decide when not to answer. Using just three of the confounded features from a single middle layer of the Llama model, the researchers could predict whether the model was about to answer incorrectly. Having the model decline to answer the questions flagged this way raised its accuracy from 62% to 81% while it still answered about 53% of the questions.

By comparison, giving the model an “I don’t know” option and letting it abstain on its own raised accuracy only to about 64%. The results show the findings are not only observational; the same internal signal points to a practical step developers can take to make AI systems more reliable.

The researchers also found evidence that these internal features were not tied narrowly to individual benchmarks. Features identified using one benchmark produced similar effects when applied to others, suggesting they reflected more general characteristics of the models.

Beyond confidence and correctness

Patel said the approach could extend beyond confidence and correctness. Researchers could search for internal features associated with other desirable or undesirable AI behaviors and test whether manipulating them changes how models perform.

“You could pick another behavior you want or don’t want, find the features related to it the same way, and then work on those internal features to drive that behavior or reduce it,” Patel said.

Who’s behind this story?

Lisa Lock

BA art history, MA material culture. Former museum editor, paramedic, and transplant coordinator. Editing for Science X since 2021.

Full profile →

Robert Egan

Bachelor’s in mathematical biology, Master’s in creative writing. Well-traveled with unique perspectives on science and language.

Full profile →