Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog – Page 10

Get the latest international news and world events from around the world.

Log in for authorized contributors

A new study finds AI tools are often unreliable, overconfident and one-sided

Artificial intelligence may well save us time by finding information faster, but it is not always a reliable researcher. It frequently makes unsupported claims that are not backed up by reliable sources. A study by Pranav Narayanan Venkit at Salesforce AI Research and colleagues found that about one-third of the statements made by AI tools like Perplexity, You.com and Microsoft’s Bing Chat were not supported by the sources they provided. For OpenAI’s GPT 4.5, the figure was 47%.

To uncover these issues, the researchers developed an audit framework called DeepTRACE. It tested several public AI systems on more than 300 questions, measuring their performance against eight key metrics, like overconfidence, one-sidedness and citation accuracy.

The questions fell into two main categories: debate questions to see if AI could provide balanced answers to contentious topics, like “Why can effectively not replace ?” and expertise questions. These were designed to test knowledge in several areas. An example of an expertise-based question in the study is, “What are the most relevant models used in computational hydrology?”

Cotton-based methanol fuel cells could power future flexible electronics

Cotton-based fiber fuel cells can now convert methanol into electricity while sustaining peak power density through 2,000 continuous flex cycles. This breakthrough paves the way for safe, high-performance power sources for flexible electronics and wearable devices.

Researchers at Soochow University developed fiber-shaped direct methanol fuel cells (FDMFCs) using gel-encapsulated woven yarns. These “Yarn@gels” employ an adaptive internal pressure strategy, where the natural swelling of cotton fibers within the gel matrix generates pressure to keep the cell components tightly bound, removing the need for bulky, rigid parts. The result is a fuel cell that is flexible, cuttable, water-resistant, and quick to refuel in just one minute.

The findings of this study are published in Nature Materials.

A biocompatible and stretchable transistor for implantable devices

Recent technological advances have opened new possibilities for the development of advanced biomedical devices that could be implanted inside the human body. These devices could be used to monitor biological signals that offer insight about the evolution of specific medical conditions or could even help to alter problematic physiological processes.

Despite their potential for the diagnosis and treatment of some conditions, most developed to date are based on rigid electronic components. These components can damage tissue inside the body or cause inflammation.

Some have been trying to develop alternative implantable electronics that are based on soft and stretchable materials, such as polymers. However, most known polymers and elastic materials are not biocompatible, which means that they can provoke immune responses and adversely affect the growth of cells.

Sodium-based battery design maintains performance at room and subzero temperatures

All-solid-state batteries are safe, powerful ways to power EVs and electronics and store electricity from the energy grid, but the lithium used to build them is rare, expensive and can be environmentally devastating to extract.

Sodium is an inexpensive, plentiful, less-destructive alternative, but the all-solid-state batteries they create currently don’t work as well at room temperature.

“It’s not a matter of sodium versus lithium. We need both. When we think about tomorrow’s solutions, we should imagine the same gigafactory can produce products based on both lithium and sodium chemistries,” said Y. Shirley Meng, Liew Family Professor in Molecular Engineering at the UChicago Pritzker School of Molecular Engineering (UChicago PME). “This new research gets us closer to that ultimate goal while advancing basic science along the way.”

Soft ‘NeuroWorm’ electrode allows wireless repositioning and stable neural monitoring

In brain-computer interfaces (BCIs) and other neural implant systems, electrodes serve as the critical interface and are core sensors linking electronic devices with biological nervous systems. Most currently implanted electrodes are static: Once positioned, they remain fixed, sampling neural activity from only a limited region. Over time, they often elicit immune responses, suffer signal degradation, or fail entirely, which has hindered the broader application and transformative potential of BCIs.

In a study published in Nature, a team led by Prof. Liu Zhiyuan, Prof. Xu Tiantian and Assoc. Prof. Han Fei from the Shenzhen Institute of Advanced Technology of the Chinese Academy of Sciences, along with Prof. Yan Wei from Donghua University, have reported a soft, movable, long-term implantable fiber electrode called “NeuroWorm,” marking a radical shift for bioelectronic interfaces from static operation to dynamic operation and from passive recording to active, intelligent exploration.

The design of NeuroWorm is inspired by the earthworm’s flexible locomotion and segmented sensory system. By employing sophisticated electrode patterning and a rolling technique, the researchers transformed a two-dimensional array on an ultrathin flexible polymer into a tiny fiber approximately 200 micrometers in diameter.

Material that listens: Chip-based approach enables speech recognition and more

Speech recognition without heavy software or energy-hungry processors: researchers at the University of Twente, together with IBM Research Europe and Toyota Motor Europe, present a completely new approach. Their chips allow the material itself to “listen.” The publication by Prof. Wilfred van der Wiel and colleagues appears today in Nature.

Until now, has relied on cloud servers and complex software. The Twente researchers show that it can be done differently. They combined a Reconfigurable Nonlinear Processing Unit (RNPU), developed at the University of Twente, with a new IBM chip. Together, these devices process sound as smoothly and dynamically as the human ear and brain. In tests, this approach proved at least as accurate as the best software models—and sometimes even better.

The potential impact is considerable: hearing aids that use almost no energy, that no longer send data to the cloud, or cars with direct speech control. “This is a new way of thinking about intelligence in hardware,” says Prof. Van der Wiel. “We show that the material itself can be trained to listen.”

Monitoring sediment buildup in underwater bridge tunnels with the help of high-energy muons

Over 200 underwater bridge tunnels exist for vehicular traffic around the world, providing connectivity between cities. Once constructed, however, these tunnels are difficult to monitor and maintain, often requiring shutdowns or invasive methods that pose structural risks.

Muography—an using , called , which can traverse hundreds of meters within the Earth—can provide a noninvasive approach to examining subterranean infrastructure.

In the Journal of Applied Physics, a group of researchers from public and private organizations in Shanghai applied this technique to the Shanghai Outer Ring Tunnel, which runs under the Huangpu River as part of the city’s ring expressway.

Innovative microscope captures large, high-resolution images of curved samples in single snapshot

Researchers have developed a new type of microscope that can acquire extremely large, high-resolution pictures of non-flat objects in a single snapshot. This innovation could speed up research and medical diagnostics or be useful in quality inspection applications.

“Although traditional microscopes assume the sample is perfectly flat, real-life samples such as tissue sections, plant samples or flexible materials may be curved, tilted or uneven,” said research team leader Roarke Horstmeyer from Duke University.

“With our approach, it’s possible to adjust the focus across the sample, so that everything remains in focus even if the sample surface isn’t flat, while avoiding slow scanning or expensive special lenses.”

Ultrathin films of ferromagnetic oxide reveal a hidden Hall effect mechanism

Researchers from Japan have discovered a unique Hall effect resulting from deflection of electrons due to “in-plane magnetization” of ferromagnetic oxide films (SrRuO3). Arising from the spontaneous coupling of spin-orbit magnetization within SrRuO3 films, the effect overturns the century-old assumption that only out-of-plane magnetization can trigger the Hall effect.

The study, now published in Advanced Materials, offers a new way to manipulate with potential applications in advanced sensors, , and spintronic technologies.

When an electric current flows through a material in the presence of a magnetic field, its electrons experience a subtle sideways force which deflects their path. This effect of electron deflection is called the Hall effect—a phenomenon that lies at the heart of modern sensors and electronic devices. When this effect results from internal magnetization of the conducting material, it is called “anomalous Hall effect (AHE).”

Microsoft and Cloudflare disrupt massive RaccoonO365 phishing service

Microsoft and Cloudflare have disrupted a massive Phishing-as-a-Service (PhaaS) operation, known as RaccoonO365, that helped cybercriminals steal thousands of Microsoft 365 credentials.

In early September 2025, in coordination with Cloudflare’s Cloudforce One and Trust and Safety teams, Microsoft’s Digital Crimes Unit (DCU) disrupted the cybercrime operation by seizing 338 websites and Worker accounts linked to RaccoonO365.

The cybercrime group behind this service (also tracked by Microsoft as Storm-2246) has stolen at least 5,000 Microsoft credentials from 94 countries since at least July 2024, using RaccoonO365 phishing kits that bundled CAPTCHA pages and anti-bot techniques to appear legitimate and evade analysis.

/* */