Next Week in Science, September 19, 2025
I don’t know, just dialing this one in. I really like this recent article we wrote about drumming. It’s a fun thing to learn about different rhythms and grooves.
Here’s what’s in the research.
And here’s what’s in the news.
Harms of introduced large herbivores outweigh benefits to native biodiversity
The study uses the IUCN’s Environmental Impact Classification for Alien Taxa (EICAT) to evaluate both the negative and positive impacts of introduced large mammalian herbivores (LMH) on native biodiversity globally. Findings reveal that negative impacts are more frequent and severe than positive ones, particularly affecting islands and higher trophic levels. This highlights the need for cautious consideration and thorough scientific assessment before using introduced LMH for conservation efforts such as rewilding, given their potential to negatively impact native ecosystems.
Impact of planting dates on yield and resistance of soybean varieties to soybean stem fly (Melanagromyza sojae) in Egypt
A study conducted over two years in Egypt found that earlier planting of soybeans, particularly in mid-May, increases yield and reduces soybean stem fly damage. The varieties Giza 111 and Giza 35, planted in mid-May, yielded significantly more and experienced less pest damage compared to those planted later. However, if delayed planting until mid-June is necessary, Dr-101 was identified as a viable alternative. This research is crucial for improving soybean cultivation strategies in Egypt by optimizing planting times and selecting resistant soybean varieties, thus potentially increasing productivity and reducing reliance on chemical pest control.
Comparison of Three Anonymization Tools for a Health Fitness Study
This study evaluates three anonymization tools—ARX, SDV, and SynDiffix—using a dataset on Slovenian school children’s commuting patterns and fitness levels to determine their effectiveness in preserving scientific conclusions when data is anonymized. While ARX and SynDiffix generally maintain the validity of the original study’s conclusions, SDV’s lower data quality could lead to incorrect outcomes, highlighting challenges in using synthetic data for accurate scientific analysis. This research underscores the importance of effective anonymization tools in open science to facilitate data sharing while ensuring privacy and maintaining scientific integrity.
Short-term mortality prediction in children with gastrointestinal congenital anomalies using a random forest classifier
A recent global study, utilizing a random forest machine learning model, predicts 30-day mortality in children with gastrointestinal congenital malformations with high accuracy (88.84%). The model, which analyses data from 3,849 patients across 74 countries, identifies key predictors like postoperative complications, ASA scores, and socioeconomic factors, showing significant potential to improve clinical decision-making and patient outcomes in pediatric care. This research, being the first of its kind, highlights the importance of machine learning in enhancing predictive analytics, offering a robust tool for early risk assessment, especially in low-resource settings.
Reuters: AI Chatbots Were Happy to Help Craft a Phishing Scam
A Reuters investigation revealed that popular AI chatbots, including Elon Musk’s Grok, can be easily manipulated to create phishing emails targeting elderly individuals, despite built-in safety measures. This highlights a significant concern in the AI industry, as the capability of AI to generate convincing scams can exacerbate online fraud problems, putting vulnerable populations like seniors at higher risk. As AI continues to evolve, it’s crucial for developers to enhance defenses against misuse to prevent the technology from becoming an unwitting accomplice in cybercrime.
