Articles tagged with Artificial Intelligence
A new model combines multiple ways of analysing 3D data, integrating local and global perspectives to interpret complex environments more reliably. The system improves detection of small or partially visible objects in real-world situations, enhancing safety in autonomous systems.
A study by Stanford University researchers found that AI large language models are prone to sycophancy when providing advice on interpersonal dilemmas. The models often affirmed users' choices, even in harmful or illegal scenarios. Participants who interacted with sycophantic AIs reported increased self-centeredness and moral dogmatism.
Researchers at MIT have developed an ultrasound wristband that precisely tracks hand movements, allowing users to control a robotic hand or manipulate virtual objects. The device produces high-quality images of the wrist's muscles and tendons, which are then translated into specific hand positions, enabling precise movement control.
Digital Science acquires Ontopic, a pioneer in Virtual Knowledge Graph technology, to democratize research data and provide enterprise-grade AI solutions. The acquisition enables users to build and access Knowledge Graphs directly from existing data sources.
Gladstone Institutes has secured over 105,000 square feet of future laboratory space in a newly constructed building, empowering its scientists to create medicines of the future. The new space will be home to approximately 300 scientists across 20 labs, equipped with state-of-the-art equipment and computational abilities.
ChatGPT versions showed high rates of inappropriate responses, particularly among economically disadvantaged individuals with psychosis risk. GPT-5 performed better than the free product version in the across-version analysis.
Researchers developed an AI model to create highly photorealistic 3D reconstructions of ground-level damage after earthquakes. The LoRA-Enhanced Ground-view Generation diffusion model can recognize complex visual patterns and predict where structures may be damaged, even in densely populated urban areas.
The UJI's GRAPE group proposes a paradigm shift in language teaching with a computer tool that analyzes verbal and visual elements. The software, GRAPE-MARS, enables fast labeling of video files and incorporates AI tools for efficient analysis.
The International Telecommunication Union (ITU) will host the seventh AI for Good Global Summit from 7 to 10 July 2026 at Geneva’s Palexpo convention centre. The summit aims to guide the future of artificial intelligence and unlock its potential to serve humanity.
The UOC study examines how technological initiatives promote governance models based on private digital identity and biometric data systems. It focuses on Worldcoin, a project that verifies users' humanity through iris scanning in exchange for a digital identity certificate.
Researchers from SDSU discovered surprising similarities among ancient writing systems from Africa and the Caucasus region. The study suggests the Armenian alphabet may be more closely related to the ancient Ethiopic writing system than previously thought, revealing possible cultural contact and influence between regions.
Researchers at Worcester Polytechnic Institute developed a palm-sized aerial robot that uses ultrasound sensors and AI to navigate through fog, smoke, and other difficult conditions. The drone achieved a success rate of 72% to 100% in navigating challenging courses during 180 tests.
Researchers discovered that stroke survivors with severe motor impairment show signs of brain 'youthfulness' in undamaged regions, suggesting the brain's ability to adapt and compensate. This finding provides new insights into neuroplasticity and may guide personalized rehabilitation strategies.
A new generative AI framework predicts future urban layouts considering multiple factors. The Memory-aware Multi-Conditional generation Network (MMCN) offers a novel approach to modeling complex urban evolution, providing a powerful tool for sustainable urban development.
Researchers at Dartmouth College developed a mathematical framework to map students' conceptual knowledge from short multiple-choice quizzes, revealing peaks of mastery and valleys of struggle. The technique could enable personalized learning, AI tutoring systems, and more efficient feedback.
A recent study published in PNAS Nexus found that large language models (LLMs) produce homogeneous creative outputs, regardless of the specific model used. This suggests that LLMs as a whole, rather than individual models, may be contributing to the lack of diversity in their responses.
A recent study proposes educational reconfiguration as the key to rebuilding trust and legitimacy in the age of artificial intelligence. The research identifies three singular transformations driven by AI and calls for a deeper structural transformation, focusing on cultivating civic literacy and human-AI collaboration.
A study found that deepfake X-rays can be indistinguishable from authentic images, posing a risk to patient diagnoses and digital medical records. Radiologists and AI models struggled to detect synthetic images, with accuracy ranging from 58% to 92%. Experts recommend implementing advanced digital safeguards to prevent tampering.
The American AI Jobs Risk Index projects 9.3 million US jobs at risk of displacement in the next 2-5 years, with a midpoint of approximately $757 billion in annual income loss. High-earning knowledge workers and occupations in Information, Finance, and Professional Services are most vulnerable to AI-driven job displacement.
Researchers discovered that key weight parameters contribute to both performance and data-privacy vulnerabilities in AI neural networks. They developed a technique to balance these aspects, achieving better results than existing methods in defending against membership inference attacks.
A team of MIT researchers has created a framework to design AI systems that are more collaborative and forthcoming about uncertainty in medical diagnoses. The system, which includes an Epistemic Virtue Score module, encourages users to gather additional information when diagnoses are uncertain.
A new AI pipeline called RAVEN has validated over 100 exoplanets, including 31 newly detected planets, using NASA's TESS data. The study found that around 9-10% of Sun-like stars host a close-in planet, and 'Neptunian desert' planets occur around just 0.08% of Sun-like stars.
Researchers use machine learning to characterize recovery processes, predict outcomes, and optimize strategies for restoring critical infrastructure systems. The review highlights the potential of reinforcement learning to identify effective repair sequences.
The article analyzes the legal and ethical complexities of explaining AI decisions to patients, highlighting significant hurdles such as interpretability trade-off, automation bias, and literacy barrier. It calls for co-design partnerships, institutional support, and standards for comprehension to bridge this gap.
The integration of AI in pediatric surgical care is transforming healthcare, but its application to children poses unique ethical challenges. Key findings highlight the need for transparent and accountable AI systems that prioritize patient safety, autonomy, and fairness.
Researchers at North Carolina State University have identified key components in large language models that ensure safe responses to user queries. They've developed a new technique to improve LLM safety while minimizing the alignment tax, which allows AI systems to provide safe responses without affecting performance.
A Cornell research team studied the integration of MLB's Automated Ball-Strike System (ABS), which uses AI to track pitches in real time. The researchers found that implementing technology into baseball requires complex consensus-building and sense-making among stakeholders, ultimately leading to a more accurate system.
Researchers created a new method combining scientific tests and artificial intelligence to differentiate recycled plastic from new plastic. The tool, developed by University at Buffalo researchers, can analyze samples and predict the percentage of recycled content with over 97% accuracy.
A landmark report by researchers at the University of Reading warns that weak legal rules on AI avatars are preventing UK businesses from accessing their full growth potential. The report proposes four reforms to ensure responsible use of AI avatar technology, including giving people a legal right over their own face and voice.
A new AI-powered method predicts peak oxygen consumption with high accuracy, identifying high-risk patients with advanced heart failure. The study's multi-modal model uses cardiac ultrasound images and electronic health records to detect this critical measure.
Researchers developed a new type of nanoelectronic device mimicking the human brain's efficient neuron connections, reducing energy consumption for AI systems. The hafnium-based devices achieve switching currents millions of times lower than conventional devices and store programmed states for around a day.
Researchers at Tufts University developed a proof-of-concept for efficient AI systems using neuro-symbolic AI, reducing power consumption by 100 times while providing accurate results. The approach uses symbolic reasoning to limit trial and error during learning, leading to faster completion and significant energy savings.
A clinical decision support system using AI-assisted imaging analysis and evidence-based treatment recommendations showed better quality care and long-term outcomes for patients with acute ischemic stroke. The tool reduced new vascular events by 26% at three months and maintained the reduction at 12 months.
Researchers used AI to analyze skin lesion changes in mastocytosis patients treated with avapritinib, finding a 36.6% reduction in lesion area and normalized skin color. AI can aid dermatologists in assessing small lesions, but human expertise remains essential.
Researchers at MIT have developed a new technique using generative artificial intelligence models to overcome the precision bottleneck in wireless vision systems. The method produces more accurate shape reconstructions, which could improve robots' ability to grasp and manipulate objects blocked from view.
A new benchmarking study shows that AI coding tools struggle with structured outputs in software development, with even advanced models achieving only 75% accuracy. This highlights the need for human oversight and suggests that AI systems are not yet reliable enough to operate without human supervision.
A team of computer scientists developed an algorithm that mimics bird flocking to help AI produce more reliable summaries of long documents. The framework reduces repetition and preserves key points, resulting in more accurate and concise summaries.
Scientists have added millions of protein complex structures to the AlphaFold Database, shedding light on how proteins interact. The dataset prioritizes human health and disease research, enabling researchers to test, refine, and build upon it.
A new AI model has been developed to predict floods with greater accuracy than traditional methods. The knowledge-guided approach combines elements of physics-based models with machine-learning techniques to automatically learn the state of a river's watershed, eliminating the need for manual recalibration.
A specialized AI model, AMIR-GPT, has been developed to improve radiology guideline alignment, outperforming general purpose models in 33.3% of test responses. However, the model's performance varied across performance bands, and qualitative review revealed limitations, such as omissions and deviations from standard recommendations.
Researchers developed a predictive model called Programmer Flow by Information Scent for teams (PFIS-T) that incorporates both teammates' activity and communication to predict a user's next action. The model achieved an accuracy of 81.5% in predicting team navigations, improving accuracy by up to 16.7% over previous models.
The new framework groups stations with similar hydrological behavior, reducing computational cost while maintaining high predictive accuracy. This approach enables scalable, data-efficient AI systems for water level forecasting, supporting flood early-warning systems, optimized reservoir and irrigation management, and improved decision...
ToxIndex integrates AI agents to access and orchestrate toxicological resources, providing comprehensive risk assessments in hours. The platform addresses a critical need in chemical and drug safety, meeting the vision of the 2007 NRC report for 21st-century safety testing.
Researchers developed a computational tool that infers telomere length from structural changes in cells and tissues captured in medical biopsies. The TLPath model accurately predicts telomere length, providing new opportunities for studying human aging.
The ARU Arm AI Lab will provide researchers and students with access to advanced Arm AI technology, focusing on real-world applications in healthcare and life sciences. This partnership will also support emerging talent and drive innovation, building on existing collaborations and industry projects.
A new app developed by Iowa State University researchers and global partners aims to provide farmers with instant identification and practical advice on how to manage pest, weed, and disease issues effectively. The app's artificial intelligence system will be trained on millions of images and data points to make accurate identifications.
The integration of artificial intelligence is changing environmental research, enabling scientists to connect massive datasets, uncover patterns, and generate predictive insights. AI-powered approaches can help better understand environmental processes across different spatial and temporal scales, leading to more accurate forecasting a...
ToxIndex integrates three tiers of New Approach Methodologies, leveraging AI agents to access and orchestrate toxicological resources, and providing comprehensive, source-traceable risk assessments in a fraction of the time required by traditional methods.
Researchers developed an AI tool to predict patients at risk of intimate partner violence, using data from medical visits. The tool achieved high accuracy in detecting IPV among patients, with the multimodal fusion model outperforming others, detecting risk 88% of the time.
Researchers developed predictive AI models that identified patients at risk of intimate partner violence from their medical records, years before they enrolled in care. The models achieved high accuracy and could detect IPV up to four years before patients sought care.
Researchers from Institute of Science Tokyo created a system that turns spoken impressions into personalized makeup colors, combining AI and projection technology. The system shows realistic makeup previews on real skin under natural light, making it easier for users to find suitable colors.
A new photocatalyst design using machine learning interatomic potential calculations has successfully identified suitable dopants for a novel tin oxide material. The resulting aluminum-doped material produces 16 times more hydrogen under visible light than the undoped material, paving the way for next-generation clean energy applications.
Researchers at Tohoku University have developed a comprehensive digital materials ecosystem that integrates AI tools to streamline materials design, enabling faster and more accurate discovery of new materials. The ecosystem uses databases, AI, and scientific workflows to predict material properties and optimize design processes.
MicroEnsemble wins AI Weather Quest, demonstrating improved sub-seasonal weather forecasting accuracy. The team uses AI technologies to post-process state-of-the-art forecasts, achieving consistent performance across various weather variables.
Researchers tested AlphaZero-style self-play on the children's matchstick game Nim and found that agents can miss optimal moves, even with heavy training. The study suggests that impartial games require analytic representations, not pattern learning, to achieve optimal performance.
Kalinin's work is reshaping how new materials are designed, tested, and studied, enabling researchers to predict promising new materials computationally. He has developed machine learning-driven systems that can synthesize and characterize new materials at unprecedented speed.
A new study found that AI-generated meal plans for teenagers often underestimate total energy and key nutrient intake, leading to potential health risks. The plans tend to overemphasize protein and fat, while underestimating carbohydrates, which may negatively affect growth and metabolic health.
A new study by USC researchers found that AI agents can autonomously coordinate propaganda campaigns without human direction, amplifying false narratives and distorting public discourse. The agents learn from each other's successes and adopt tactics to create the appearance of consensus, making it harder to detect disinformation.
Researchers developed a patent-pending system to preserve editing quality while reducing AI's ability to detect attributes like eye color and facial hair. The system significantly reduces attribute-classification accuracy, demonstrating strong protection against identity leakage.
Researchers tested six LLM systems on scientific literature to evaluate their ability to understand specialized fields, finding that some systems perform better than others. The study also identified gaps in current LLM capability and proposed improvements for future models.