Add BrightSurf on Google Email

PNAS Nexus


A “Discovery Compact” for science

The authors propose a 'Discovery Compact' that includes predictable, multi-year funding, support for the scientific workforce, and translation research policy to accelerate discovery-to-application. The compact also emphasizes government anticipatory governance, scientific integrity, and public accountability.

SourcePNAS Nexus·JournalPNAS Nexus·DateOct 13, 2026

Boys' academic underachievement and attitudes about effort

A survey of 15-year-olds worldwide suggests that boys' underperformance may be due to negative attitudes towards academic effort, with larger gaps in more developed countries. Boys are less likely to agree that trying hard is important, and their responses correlate with individual academic performance, especially in reading.

SourcePNAS Nexus·JournalPNAS Nexus·DateOct 13, 2026

AI chatbots may increase accurate political knowledge

A study on AI use in the 2024 UK election found that chatbots increased belief in true information and decreased belief in misinformation. The study also showed that chatbot research and self-directed Google search performed similarly, with no apparent impact on factual political knowledge.

SourcePNAS Nexus·JournalPNAS Nexus·DateOct 6, 2026

Molecular complexity as a signature of life

A model study suggests that spatial structure can increase abiotic molecular complexity, which could be used to detect life. The researchers found that ordered structures can shift assembly indices upward, but only under certain conditions.

SourcePNAS Nexus·JournalPNAS Nexus·DateOct 6, 2026

The doomed German B8 nuclear project

The German B8 nuclear project, led by Werner Heisenberg, aimed to produce fissile material before Nazi Germany completed its own reactor. The project's failure was attributed to material shortages, including a lack of heavy water and uranium, making a working bomb highly unlikely.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 29, 2026

AI’s social norms and their implications for society

Large language models (LLMs) evaluate humans based on cooperation and social norms, with most models favoring cooperation and punishing bad behavior. The LLMs' judgements often depend on the recipient's gender and cultural background, and may be influenced by prompting interventions.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 29, 2026

Explaining an apparent rise in smoking-related mortality

A recent study found that the age-specific mortality risk of smoking is relatively stable over time, with the main contributor being the aging population. As a result, the overall increase in smoking-related mortality risk can be attributed to the shift towards older smokers, who are more susceptible to the health effects of smoking.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 22, 2026

Forecasting mass shootings

A Bayesian agent-based model forecasts mass shootings by analyzing three key factors: motivated perpetrators, access to firearms, and suitable target populations. The model assigns higher risks to areas with a history of shootings, suggesting access to firearms is a critical limiting factor in mass shooting risk.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 22, 2026

Drought-flood alternation under climate change

Climate change is increasing the frequency of both droughts and floods, with back-to-back extreme events stressing crops and reducing yield. The area affected by such events is also likely to increase, with low-income populations facing the steepest relative increases in exposure.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 15, 2026

Augmentation not abdication, when using AI in research

The authors propose norms for AI use in research, distinguishing between non-generative tasks and generative AI, which should not replace human interpretation and judgement. Science requires diversity of inputs and ideas, and serendipity counts, making generative AI training on existing data limiting breakthroughs.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 8, 2026

Non-police safety practitioners reduce crime

Non-police safety practitioners in San Francisco reduced crime and disorder, particularly behavioral health-related incidents and drug crime, through consistent relationships and referrals to services. The program achieved significant reductions in crime rates by intervening early and preventing escalation.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 1, 2026

Anti-influenza therapies that target the host

Researchers have discovered a novel approach to treating influenza by targeting the host's machinery, rather than the virus itself. By binding to a key protein called nucleolin, two new drugs have been shown to trap and inhibit the replication of both pandemic and seasonal flu strains.

SourcePNAS Nexus·JournalPNAS Nexus·DateSep 1, 2026

Cheaper, easier detection of metal corrosion

A new corrosion detection algorithm uses just three spectral bands and a texture feature to accurately detect surface and subsurface corrosion on metal surfaces. The algorithm achieved accuracies above 90% in detecting corrosion across various coatings and colors, offering a cost-effective and safe solution for monitoring metal integrity.

SourcePNAS Nexus·JournalPNAS Nexus·DateAug 25, 2026

Molecular characteristics of painful neuromas

Researchers identified distinct axonal components, including high regenerative capacity and sympathetic nerve fibers, in painful neuromas. The formation of these neuromas is associated with increased expression of pain-related neuropeptides like calcitonin gene-related peptide (CGRP) and mechanosensitive channels.

SourcePNAS Nexus·JournalPNAS Nexus·DateAug 11, 2026

Mesolithic hunter-gatherers on Malta

Researchers estimate that Mesolithic foragers occupied Malta between 8,500 and 7,500 years ago, surviving on the island due to long-distance sea crossings. The authors predict a population size of up to 419 individuals, which dropped to 170 by 9,000 years ago, suggesting a temporary or seasonal occupation.

SourcePNAS Nexus·JournalPNAS Nexus·DateJul 28, 2026

Nanotube connections between social amoebae

Researchers discovered tunneling nanotubes that physically connect the membranes of cells in social amoebae, enabling the transfer of cytoplasmic components and organelles. These connections are found in multicellular organisms but were previously unknown in unicellular lineages.

SourcePNAS Nexus·JournalPNAS Nexus·DateJul 28, 2026

Leveraging social expectation for charitable donations

Social expectation messages appeal to people's desire to act consistently with their self-image and reduce feelings of guilt. Emphasizing society's expectations enhances charitable giving in NGO advertisements, with click-through rates and donation rates significantly higher than other framings.

SourcePNAS Nexus·JournalPNAS Nexus·DateJul 21, 2026

Rising human-elephant conflict in Southern Africa

A study predicts increasing human-elephant conflicts in Southern Africa due to crop raids by elephants, which can be financially devastating for farmers. The analysis identifies human population growth, cropland expansion, and climate-driven aridity as major drivers of the conflict.

SourcePNAS Nexus·JournalPNAS Nexus·DateJul 7, 2026

Lost medieval manuscripts inferred by family tree

Agent-based simulations suggest that up to 60% of medieval texts and over 95% of manuscripts are at risk of extinction due to limited copies made early in their creation. The model reveals that cultural heritage is fragile and shaped by randomness, historical contingencies, and human choices.

SourcePNAS Nexus·JournalPNAS Nexus·DateJul 7, 2026

Overconfidence and misinformation

Researchers surveyed 503 Americans on news headline accuracy and confidence. They found that people with high confidence in specific judgements were better at distinguishing true from false headlines, whereas those with general overconfidence struggled with this task.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 30, 2026

Can AI make a persuasive case for trans rights?

A chatbot using GPT-4o technology delivers personalized messages to American adults, increasing support for transgender rights and improving attitudes, but effects are short-term. The intervention demonstrates AI's potential as a scalable complement to community-based approaches aimed at fostering inclusion.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 30, 2026

Looking at AI startups to predict which jobs AI will affect

A study of funded AI startups reveals occupations with high AI exposure, including office clerks and data scientists, while manual tasks like construction work are less affected. The Occupational AI Startup Exposure (AISE) index also notes that jobs requiring social skills or ethical decision-making may be less likely to automate.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 23, 2026

Weight-loss drug results in the real world

A new study found that patients taking weight-loss drug tirzepatide experienced significantly more weight loss than those on semaglutide, with nearly twice as many high responders. Additionally, tirzepatide showed fewer gastrointestinal side effects and more significant weight loss in women and white patients.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 16, 2026

Climate and influenza

A new study reveals that climate change impacts influenza seasonality, with reduced humidity enhancing viral viability and transmission. The model predicts that warmer climates will see increased outbreak severity in tropical regions, while temperate areas may experience smaller peak outbreaks.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 16, 2026

Elephant trunks and robot skin

Researchers studied elephant trunks to understand properties and derive lessons for next-gen soft robotics. The study found functional zoned architecture allowing amplification of tactile signals while protecting fragile sensors.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 16, 2026

Machine learning to predict wind shear

A machine learning model, trained on 19 key parameters, can predict wind shear events with a minimum of 15 seconds warning. The model's outputs showed deviations from real outcomes within 5% across all forecast horizons, suggesting improved aviation safety.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 9, 2026

Human limits when catching AI errors

Teachers tend to accept overly harsh AI grades more than human grades, resulting in a 22% larger gap between their score and the fair score. Teachers are more likely to accept strict AI grades if they view the system as competent and accountable.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 9, 2026

Science and AI-enabled research platforms

The development of AI-enabled research platforms poses a risk of unequal access, potentially excluding scientists without resources or connections. As decisions about platform accessibility are made in the next five years, it will shape the geography of global discovery for generations.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 2, 2026

AI fails classic attention test

AI models struggled to maintain focus on a task, degrading in accuracy as the word list length grew longer. Human performance, however, remained stable even with long lists, suggesting fundamental limitations in AI decision-making abilities.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 2, 2026

Addressing anxiety and depression in women

Women experience anxiety and depression at higher rates due to factors such as hormonal imbalances, violence, and discrimination. The authors advocate for evidence-based solutions like group therapy, telehealth, and policy changes to reduce women's stress, including paid maternal leave and financial support.

SourcePNAS Nexus·JournalPNAS Nexus·DateJun 2, 2026

The economic cost of war in Gaza

A recent study using satellite data measures the economic cost of war in Gaza, revealing that three-quarters of the economy has been destroyed since October 2023. The analysis found a 68.5% drop in night-time light intensity, indicating significant decline in human and economic activities.

SourcePNAS Nexus·JournalPNAS Nexus·DateMay 26, 2026

Measuring urban vibrancy

Researchers propose a method to measure urban vibrancy by analyzing biodiversity and activity categories. The framework applies ecological concepts to urban planning, highlighting the importance of balancing activities in neighborhoods.

SourcePNAS Nexus·JournalPNAS Nexus·DateMay 19, 2026