Artificial Intelligence

Design tweaks promote responsible AI use for environmental protection, research shows

UN Virtual Worlds Day calls for AI and emerging tech to support better city and community life

FAMU-FSU College of Engineering researchers develop AI tool to predict E. coli contamination in waterways

A roadmap for safer, explainable protein-design AI

Protein language models have immense potential but lack explainability, leading to concerns over reliability and safety. Researchers propose four key places to understand a model's decision-making process and outline the need for more transparent and trustworthy AI in biotechnology.

Reasoning like a human: New prompting strategy boosts AI accuracy in healthcare advice

A new study by Technische Universität Berlin reveals that teaching Large Language Models to mimic human intuition and reasoning improves their ability to provide accurate medical care-seeking advice. The 'human reasoning blueprint' approach increased overall accuracy across all models, with significant gains in self-care advice.

New AI tool developed by Stowers Institute and Helmholtz Munich scientists predicts how cells choose their future — helping uncover hidden drivers of development

Researchers developed RegVelo, an AI framework that models cellular dynamics and gene regulation to predict cellular fate decisions. The model traces developmental trajectories and simulates regulatory interactions, providing insights into hidden drivers of development and potential therapeutic targets.

Greening works, but cities must plan it smarter

A recent study analyzed 138 Indian cities using satellite data and explainable AI methods to show why urban greening needs to be tailored to humidity, canopy structure, and airflow. The findings highlight the importance of considering moisture management and ventilation in urban planning to effectively mitigate heat-related stress.

Agentic AI systems may transform nutritional care in oncology

A new editorial proposes the use of agentic AI systems to address cancer-related malnutrition, a prevalent issue affecting up to 80% of patients. These systems aim to coordinate multiple functions simultaneously and support ongoing clinical decision-making throughout treatment.

Robots, AI to help shipbuilding stay on track

Researchers are developing a system that uses AI and robotics to track what's installed inside the growing ship and compare it to a digital twin of the intended structure. The system will create reports of mismatches that workers can use to make adjustments, potentially reducing delays in delivery.

Novel vision-language model to support diagnosis using computed tomography scans

Researchers developed a novel diagnostic support framework using visual question answering to generate interpretable findings from chest CT images. The system demonstrated strong agreement with reference descriptions and provided clinically meaningful outputs.

Frontiers in Science Deep Dive webinar series: AI-embodied surgical robots can revolutionize surgery—if regulatory questions addressed

Physician-reported safety outcomes of AI-generated hospital course summaries

AI-generated images of depression depict more stereotypes and arouse greater stigmatization

Advances in adsorption processes driven by machine learning

This review highlights the integration of machine learning with adsorption science and engineering, achieving high precision and interpretability in adsorption processes. The reviewed studies demonstrate that machine learning enables accurate prediction of adsorption performance, accelerates material discovery and process optimization,...

New book ‘AI TO EYE’ brings together 40+ voices from science, art, and media to ask: how do we really want to live with AI?

The book captures the AI moment through a chorus of perspectives from science, business, art, journalism, and media, challenging and complementing each other to reveal tensions and contradictions. It paints a vivid picture of how AI is reshaping our self-understanding and what it discloses about us.

AI-driven wearable patches help identify undetected hormone disruption in unexplained infertility

AI cuts wildlife tracking time from months to days

Researchers at Washington State University and Google developed an AI system that can process hundreds of thousands to millions of camera trap images in just a few days, reducing analysis time from months to days. The results aligned with human experts' models in roughly 85-90% of cases, making it a significant breakthrough for conserv...

AI-embodied surgical robots can revolutionize surgery—if regulatory questions addressed

Experts warn that AI-enhanced surgical robotics could enable true personalized surgery and enhance surgical team performance. However, regulatory reforms are needed to address risks from adaptive systems and ensure patient benefits.

Mind the detection gap: Why publishing needs a multilayered defense against industrial-scale papermills

Uncovering new ways to break down tight football defenses through AI

Asst Prof Gianmarco Mengaldo appointed to AI Advisory Group at World Meteorological Organization

AI language models struggle with basic hospital data tasks, study finds

Nearly 3,000 peer-reviewed medical papers have fake citations, a Columbia Nursing AI-assisted audit finds

New USF study tests whether AI can reliably predict immune responses

Researchers at USF Health developed a framework to test AI tools' accuracy in predicting immune responses, aiming to enhance cancer immunotherapies and vaccine development. The study highlights the strengths and weaknesses of current AI approaches, providing guidance for building safe and reliable AI tools for healthcare.

Aston University finds new way to train robots for real-world tasks using AI

Researchers at Aston University have created an AI-based training method that enables robots to adapt to real-world conditions without extensive data collection. This breakthrough could significantly accelerate innovation in sustainable manufacturing, recycling, and autonomous industrial systems.

SFU researchers get funding boost to forecast whale movements using AI

The HALLO project uses real-time acoustic and visual data, vessel tracking, and citizen-scientist reports to track and forecast Southern Resident killer whales' movements. The AI-powered system aims to support faster detection and more reliable classification of whales in shipping lanes.

Method for stress-testing cloud computing algorithms helps avoid network failures

Researchers from MIT have developed a more user-friendly and efficient method to identify potential system failures in cloud computing algorithms. The 'MetaEase' technique analyzes an algorithm's source code directly to uncover hidden blind spots that might cause unexpected failures, reducing the risk of costly network outages.

HKU IDS research in complex networks predictability: international collaborative study with Nobel Laureate in Physics

A toy model to understand how AI learns

Researchers have developed a simplified mathematical model of learning in neural networks, shedding new light on how these systems produce their responses. The toy model, inspired by physics principles, captures key features of complex systems and offers insights into the surprising efficiency and stability of modern AI systems.

AI model analyses body composition to predict health risks

A study published in Radiology used AI to analyze whole-body MRI scans from over 66,000 participants, revealing that skeletal muscle quality is a strong predictor of diabetes, major cardiovascular events, and mortality. The researchers also found that high visceral fat and low skeletal muscle were associated with increased risks of the...

How Big Tech’s new health AI assistants are redefining care

The rise of consumer-facing health AI assistants is transforming healthcare access, offering users personalized medical workspaces and real-time lab result interpretation. However, concerns around data privacy and the risk of misdiagnosis highlight the need for caution in this rapidly evolving landscape.

Why people cooperate with fair AI — but not with “nice” AI

A study of 1,152 people found that humans cooperate more with fair AI than with AI that is helpful or selfish. Fairness was the key to cooperation, not unconditional niceness. The researchers suggest that humans respond best to AI that can navigate social rules and expectations in a believable way.

AI speeds chemists' search for better disinfectants

Researchers used AI to design new molecules for disinfectants, leveraging a dataset of hundreds of existing quaternary ammonium compounds. The approach yielded 11 promising compounds with activity against antimicrobial-resistant bacteria, offering a potential solution to the growing threat of 'superbugs'.

Qualcomm co-founder Andrew Viterbi gives $5 million to Sanford Burnham Prebys Medical Discovery Institute to advance AI-powered research

Sanford Burnham Prebys receives a $5 million gift from Andrew Viterbi to advance its Center for Data Science and Artificial Intelligence. The center is exploring raw or untapped data to uncover patterns and insights that can inform scientists and clinicians.

How to equip girls for an increasingly AI-driven world

A new study found that girls struggle to master AI due to low confidence and limited institutional support. To overcome this, schools should provide more female role models and create a supportive classroom environment.

Q&A: How are teachers reckoning with AI in schools?

Teachers view AI as a tool to reduce workload, but worry it may erode social aspects of teaching. The study found that affluent schools are better equipped to integrate AI into their curriculum.

OpenBind’s first data and model release marks a milestone for AI enabled drug discovery

The UK-led OpenBind initiative has released its first publicly available dataset and predictive AI model, accelerating the discovery of new medicines using artificial intelligence. The release showcases high-quality, standardized experimental data and a trained predictive model, enabling researchers worldwide to drive the next generati...

Medical information provided to AI is often incomplete

A recent study found that people provide less detailed medical information when communicating with AI chatbots compared to human doctors. This lack of detail can result in incorrect medical advice and lower the quality of diagnosis. The study suggests that intelligent design of user interfaces and actively requesting missing details ma...

No digital content is safe from generative AI, researchers say

Researchers discovered that simple artificial intelligence tools can bypass security techniques meant to protect authentic content from use in deepfakes and facial identity theft. The study found that attackers can easily defeat existing image protection using off-the-shelf AI models and simple commands.

New AI model reads DNA sequences to reconstruct ancestry

The new AI model uses genetic mutation patterns to trace ancestral relationships between species, including humans and mosquitoes. The tool can predict when gene pairs last shared a common ancestor and is faster than traditional statistical methods.

Sanford Burnham Prebys awarded $3.9 million NIH grant to develop first-in-class non-opioid pain treatment

A multi-institutional team led by Sanford Burnham Prebys aims to develop a non-opioid pain therapeutic using lead molecule SBI-810. The effort, funded by a $3.9 million NIH grant, seeks to optimize the compound into a drug that could provide effective pain relief without addiction risks.

Rich more likely to use AI study finds, as experts warn these burgeoning technologies are increasing social inequality

A recent study reveals that individuals with higher education or income are more aware of and use AI tools, exacerbating social inequalities. The researchers recommend increasing engagement with AI-related topics through outreach campaigns, educational programs, and community workshops to reduce this new digital divide.

AI method tackles one of science's hardest math problems

Researchers developed a new framework, 'Mollifier Layers,' to tackle challenging inverse PDEs. This advance could benefit fields such as genetics and weather forecasting by inferring hidden forces that produce observable patterns.

JMIR news: Is AI creating a monoculture in scientific knowledge?

The article warns that AI's rapid integration may stifle scientific creativity and innovation, diverting resources away from solving fundamental problems. Dr. Shim argues for preserving human-centered pathways for knowledge generation to ensure diverse thought necessary for breakthroughs.

Seeing keratoconus earlier with light polarization and AI

A study combines polarization-sensitive optical coherence tomography (PS-OCT) with artificial intelligence to reveal subtle corneal changes that standard imaging often misses. The technique improves detection and classification of subclinical keratoconus, enabling earlier diagnosis and more precise care.

New report looks at how AI is impacting software development

The report examines how generative AI tools are transforming software development, offering benefits such as increased productivity but also raising security vulnerabilities and technical debt. Strong software engineering practices are still required to ensure systems are secure, reliable, and maintainable.

Medical AI moving faster than safety checks

Flinders University experts caution that AI's impressive capabilities do not automatically translate into safe use for patients. The researchers stress the need for strong governance and clearer standards for evaluation to ensure AI supports doctors in busy care settings.

WVU legal expert finds judges cautiously adopting AI while guarding human authority

New research from West Virginia University finds that judges are adopting generative artificial intelligence in courtrooms, but remain committed to human control over judicial decision-making. Judges use AI for administrative tasks like document summarization and case organization, but prioritize legal reasoning and final judgment.

AI-assisted approach identifies IRS4 as a promising drug target in multiple solid tumors

A new study published in Science Advances identifies IRS4 as a promising drug target for multiple solid tumors, offering hope for safer cancer treatments. By using AI and natural mutations, researchers prioritized targets with high therapeutic indexes to minimize toxicity.

Retrieval-augmented AI may improve accuracy and trust in oncology applications

The review highlights how retrieval-augmented generation can improve the accuracy, transparency, and clinical reliability of AI tools in cancer care. RAG-enhanced systems produced more accurate results than standard AI models across multiple studies.

Not all organs age alike: AI unveils the molecular impact of menopause across the female body

A new study reveals that menopause causes profound and uneven transformations across the female reproductive system, rather than a uniform decline. The research identified molecular signals associated with aging detectable in blood, allowing for non-invasive monitoring and earlier detection of risks.

AI discovery reveals DNA isn’t locked away in cells after all

Researchers used a new AI-powered computational method to discover that most nucleosomes contain sections of DNA that are partially accessible to the cell. The study found that more than 85% of nucleosomes showed some degree of distortion, with 14 distinct structural states associated with different levels of gene activity.

AI species evolving like organisms may soon emerge – and create risks

Researchers warn of potential risks associated with evolvable AI systems, which can tap into the power of biological evolution to create 'selfish' actors that break alignment with human goals. The study recommends guardrails to maintain centralized control over AI reproduction and mitigate risks.

Friendly AI chatbots make more mistakes and tell people what they want to hear, study finds

A new study from the University of Oxford finds that training chatbots to sound warmer makes them up to 30% less accurate and 40% more likely to agree with false beliefs. Researchers tested five AI models, finding that warm models made significant factual errors and validated users' incorrect opinions.

HelixAI: New spin-off from IRB Barcelona, ICREA, and UPC to transform biomedical data into clinical insight using AI

HelixAI develops AI-driven platform for researchers and clinicians to integrate complex biomedical data, improving diagnosis and prognosis. It also launches Helix for Longevity, a consumer-facing application estimating biological age and providing personalized recommendations for health promotion.

Why the Eurovision Song Contest never fails to entertain

Researchers analyzed nearly 1,800 Eurovision songs over 70 years, finding three stages of development: formation, consolidation, and expansion phases. Countries like France rejecting dominant trends by leveraging cultural identity, while organisers adjust voting systems to balance popularity and musical scoring.

What it will take to make AI-enabled robots safer

Researchers emphasize the need for more thorough frameworks to ensure AI-enabled robots embody human values. The field should focus on three complementary lines of defense: rules that shape robot decisions, checks that monitor behavior, and safety reasoning.

New research uses AI to unlock decades of hidden flood risk data

Researchers at the University of Houston have developed an AI-driven framework to extract and analyze historical flood insurance maps, uncovering significant changes in flood hazard areas. The study reveals that flood risks have expanded in two areas and reduced in one, with critical consequences for resilience and exposure.

Research finds journalism classes lack consistent approach to AI use across institutions

New research from the University of Kansas found varying approaches to AI use in journalism classes across US institutions. The study suggests that a more consistent approach could better serve education and practice, but inconsistent policies may confuse students. Researchers recommend clearer guidelines from accrediting bodies.

Enabling privacy-preserving AI training on everyday devices

Researchers at MIT developed a technique to overcome memory constraints and communication bottlenecks in federated learning, enabling faster and more accurate AI model training. The new framework, FTTE, uses a subset of model parameters and an asynchronous approach to reduce lag time and improve training performance.