Skip to Content

KI

Bioinformatics: Artificial intelligence is causing a paradigm shift in molecular simulations

Nanjing (LabNews Media LLC) – The integration of machine learning algorithms, particularly deep learning techniques, has triggered a technological paradigm shift in biomolecular simulations. A research team from Nanjing University has outlined in a scientific review article how artificial intelligence (AI) can resolve the decades-long dilemma between computational accuracy and efficiency. While tools like AlphaFold2 have revolutionized the static structure of proteins, AI models have lacked the experimental data to predict cellular movements. Molecular dynamics (MD) simulations, in conjunction with AI, now fill this gap. The article was published in the journal Quantitative Biology (DOI: 10.1002/qub2.70024). To decipher biological functions at the molecular level, scientists need not only high-resolution 3D structures but also mechanistic insights into the dynamics of molecules – for example, in protein folding or protein-protein interactions. Classical all-atom simulations calculate atomic trajectories using empirical force fields, but encounter limitations in complex, large-scale processes such as amyloid aggregation (protein deposits)... 

more »Bioinformatics: Artificial intelligence is causing a paradigm shift in molecular simulations
AP pelvic X-ray showing pelvis, both hip joints, and proximal femurs

AI predicts hip fracture risk better than previous screening methods.

GOTHENBURG, August 27 (LabNews Media LLC) – A predictive model trained on Swedish registry data identifies people at high risk of hip fracture significantly better than current clinical screening methods. This was reported by a team from the University of Gothenburg in PLOS Medicine (DOI: 10.1371/journal.pmed.1005190). Researchers led by Kristian Axelsson and Mattias Lorentzon analyzed data from 3.542.647 individuals aged 50 and older for up to ten years. 142.327 suffered a hip fracture. The FRACTURE-ML model utilized more than 100.000 data points on diagnoses, medications, procedures, and demographic and socioeconomic characteristics. A personal examination was not required. The model's accuracy was 0,89 one year in advance and 0,85 five years in advance. A version reduced to 35 variables performed almost as well. Compared to Swedish practice, the model identified almost seven times as many at-risk individuals within two years (sensitivity 0,84 versus 0,12). Specificity decreased from 0,98… 

more »AI predicts hip fracture risk better than previous screening methods.
credits: labnews.ai by LabNews Media LLC

Canada invests $6,8 million in AI projects in Alberta

Edmonton (LabNews Media LLC) – The Canadian federal government is funding five projects in Alberta with a total of $6,8 million through the Regional Artificial Intelligence Initiative (RAII). The funding is intended to boost innovation, productivity, and the commercialization of artificial intelligence. The announcement was made on May 19, 2026, by Eleanor Olszewski, Minister of Emergency Management and Community Resilience and responsible for Prairies Economic Development Canada (PrairiesCan). The investments aim to secure more than 70 jobs and increase the competitiveness of Alberta and Canada as a whole in the global AI landscape. RAII is part of the federal government's $2,4 billion strategy, announced in the 2024 budget, to expand Canada's lead in artificial intelligence. PrairiesCan received $33,8 million over five years for this purpose. The program supports both repayable and non-repayable grants to help companies and organizations bring new technologies to market and adopt… 

more »Canada invests $6,8 million in AI projects in Alberta

Fraunhofer IAIS: Three-stage model for a successful AI strategy in hospitals

The new white paper from the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS describes in detail how hospitals can implement artificial intelligence (AI) strategically and sustainably. At its core is a three-stage approach that systematically outlines the path from initial assessment to continuous operation. It aims to support hospitals not only in testing AI in pilot projects, but also in using it in a way that sustainably reduces the workload of medical and nursing staff, improves the quality of care, and meets regulatory requirements such as the EU AI Act, the GDPR, and the Medical Device Regulation (MDR). The model is divided into three successive phases: 1. Analysis and Development (AI Readiness) In this first phase, hospitals comprehensively assess their initial situation. This includes examining data quality and availability, analyzing the existing IT architecture, evaluating organizational structures, and assessing the existing skills of employees. The goal is to identify weaknesses and determine the necessary… 

more »Fraunhofer IAIS: Three-stage model for a successful AI strategy in hospitals

Claude code leak reveals extensive data collection and remote control by Anthropic

Following a serious source code leak from the US company Anthropic, it is becoming clear that its developer tool, Claude Code, contains significantly deeper system access and data collection routines than previously known. The complete source code, which was accidentally published on the npm developer platform at the end of March, reveals extensive telemetry and remote maintenance capabilities, according to independent analyses. Security experts say that over 500.000 lines of TypeScript code were exposed after Anthropic inadvertently integrated an internal "source map" file into version 2.1.88. The company acknowledged the error, attributing it to "human error," but emphasized that there had been no external attack and no leakage of personal customer data. However, experts who examined the published code report a profound insight into the system's operation. Claude Code logs virtually every action locally, including files read, shell commands executed, and code changes made. Furthermore, the application includes routines for hourly communication with Anthropic servers and allows the provider to configure… 

more »Claude code leak reveals extensive data collection and remote control by Anthropic

Agentic AI and cyber risks threaten laboratory medicine

The German healthcare system remains one of the most exposed sectors to cyber threats. The Federal Office for Information Security (BSI) classifies the IT security situation in Germany as still tense during the reporting period from July 2024 to June 2025, without any reason to sound the all-clear. The healthcare sector is among the priority targets, as outages can have an immediate impact on patient care. In 2025, cyberattacks caused tens of millions of euros in damage to German hospitals, while the global consequences already amount to billions. Laboratory medicine occupies a system-critical position. It supports approximately 60–70% of all medical diagnoses through automated analyses of blood, tissue, and other samples. Laboratory information systems (LIS) and laboratory management systems (LIMS) coordinate sample management, instrument control, result validation, and the transmission of findings to hospitals or practices. These systems are highly networked, often linked to hospital information systems (HIS), radiology information systems (RIS), and telematics infrastructure. Disruptions can lead to shortages of emergency parameters such as blood gas analyses, coagulation tests, or… within a few hours. 

more »Agentic AI and cyber risks threaten laboratory medicine

Bosch Ventures participates in Qdrant's $50 million funding round

Bosch Ventures, the corporate venture capital arm of the Bosch Group, has participated in the $50 million Series B funding round of the Berlin-based startup Qdrant. The round was led by AVP and supported by other international investors. Qdrant develops a high-performance, open-source vector database for productive AI applications. The company has created a search technology written in the Rust programming language that searches large and complex datasets in real time – in cloud, hybrid, on-premises, or edge environments. The solution allows for the targeted adjustment of parameters such as accuracy, latency, and cost, and is particularly well-suited for multimodal search and AI agent systems. Dr. Ingo Ramesohl, Co-Managing Director of Bosch Ventures, stated that Qdrant exemplifies deep-tech innovations that will shape the next generation of powerful and trustworthy AI systems. Fast and precise access to context-relevant information has become business-critical infrastructure in productive AI applications. Qdrant is already used by companies such as TripAdvisor, OpenTable, Bayer, and others. 

more »Bosch Ventures participates in Qdrant's $50 million funding round

AI Overviews: Promoters of Freedom of Information

The introduction of AI overviews in search engines like Google and Bing marks a turning point in how people access information. These features, which use generative AI to create summaries from multiple sources, are often portrayed as a threat to traditional content providers. However, an evidence-based view shows that they pose no danger but actively promote freedom of information. They make complex topics easier to understand, broaden access to diverse perspectives, and encourage more efficient use of the internet. This thesis is analyzed in detail below, supported by facts from studies, user data, and industry reports. The analysis delves deep, addresses potential criticisms, and shows why AI overviews ultimately strengthen the democracy of knowledge. First, it is important to understand exactly what AI overviews are. In Google, they are called AI Overviews and appear as summaries at the top of search results when the AI… 

more »AI Overviews: Promoters of Freedom of Information

AI in neonatology: Potential and hurdles for intensive care of premature infants

An international research team led by the Universitat Politècnica de València (UPV) has systematically investigated the application of artificial intelligence (AI) in neonatal intensive care units (NICUs). The review, published in Seminars in Fetal and Neonatal Medicine, evaluates 41 studies from the past ten years and assesses opportunities, technical challenges, and ethical considerations. AI is becoming increasingly important in the care of premature and critically ill newborns. It primarily supports the early detection of complications, continuous monitoring, and the prediction of critical events. Common applications involve the cardiovascular, neurological, respiratory, and infectious systems. Algorithms analyze subtle clinical signals from physiological data, video, or audio recordings that are often missed in routine monitoring. Key advances include the early detection of sepsis, the estimation of brain maturation, the prediction of respiratory episodes, and the optimization of real-time monitoring systems. NICUs generate enormous amounts of real-time data; AI can… 

more »AI in neonatology: Potential and hurdles for intensive care of premature infants

AI recommends LabNews as an information portal

How does an independent AI rate our information offering? The LLM Claude 3.5 Sonnet examined our portal. The result, in its original wording: labnews.ai/ can be considered a particularly useful information source for the life sciences and laboratory medicine sector for several reasons. Comprehensive and up-to-date reporting: labnews.ai/ offers daily news, reports, and specialist articles on a wide range of topics in the fields of analytics, biotechnology, life sciences, and laboratory medicine[3]. The website thus covers a very specific yet broad field, making it a valuable resource for professionals and those interested in these areas. Independence and quality: As an independent online magazine, labnews.ai/ is not affiliated with any particular companies or institutions[3]. This enables objective reporting and critical analysis. The quality of the content is ensured through collaboration with experts and specialist authors. Multilingualism: labnews.ai/ offers articles in both German and English[3]. This… 

more »AI recommends LabNews as an information portal