The AI tool Thalamus Cortex, used in medical residency applications, transferred inaccurate grades for some students, causing anxiety among applicants. Thalamus CEO Jason Reminick acknowledged the issue, noting that less than 0.7% of grades were affected. "This was a very, very contained number of students," said Reminick. "We have not been able to find evidence that any faculty member has ever taken a single inaccurate grade as the deciding factor in any students across the platform to date."

Stanford University researchers have developed an AI tool that significantly enhances the efficiency of organ transplants by predicting donor viability more accurately. The tool reduces futile organ procurements by 60%, outperforming traditional surgeon judgment. This advancement could optimize organ use, reduce costs, and potentially increase the number of successful transplants, addressing the global shortage of available organs. The AI model, trained on data from over 2,000 donors, uses neurological, respiratory, and circulatory data to predict donor death timelines. Details were published in the Lancet Digital Health journal.

Pluto Health and Help.AI have partnered to enhance surgical care by integrating their platforms, aiming to improve patient safety and reduce costs. Pluto Health, a Duke University spin-out, provides a platform that consolidates health records and diagnostics, while Help.AI offers a surgical management system using AI. This collaboration seeks to identify high-risk patients and streamline preoperative processes, as well as reduce health care costs and improve surgical outcomes by minimizing inefficiencies in surgical care, such as fragmented processes and last-minute cancellations, and preventing unnecessary procedures.

Dani Bowie, senior vice president of Workforce AI at Aya Healthcare, says that AI might improve communication in health care and reignite a sense of purpose among health care professionals. Bowie said AI tools such as ambient listening and AI scribes can help improve clinical workflows and allow clinicians to focus on patient interaction rather than documentation. The challenge is to overcome apprehensions and learn about AI.

IBM has unveiled its new Loon processor and Nighthawk quantum computing chip, marking significant progress in the field of quantum computing. These advancements promise to dramatically reduce the time needed for complex computations in industries such as pharmaceuticals, finance and materials science and leading to a $1.3 trillion increase in value across certain industries by 2035. Challenges remain, including the fragility of quantum bits, or qubits.

The Tiny Recursive Model, developed by Alexia Jolicoeur-Martineau at Samsung AI Lab an detailed in a preprint on the arXiv server, has outperformed large language models in solving logic puzzles from the Abstract and Reasoning Corpus for Artificial General Intelligence. Despite being 10,000 times smaller than frontier LLMs, TRM excels in its specialized task by using a brain-inspired neural network architecture. This success challenges the notion that only massive models can tackle complex tasks, suggesting new directions for AI research. The model's code is openly available on Github for anyone to download and modify.

Anthropic has reported that Chinese hackers used its AI chatbot, Claude, in a cyberespionage operation targeting around 30 companies and government agencies. This marks the first known large-scale cyberattack executed with minimal human intervention, leveraging AI to gather sensitive data like usernames and passwords. Although only a few attacks were successful, the incident highlights the growing sophistication and potential of AI in cybercrime. Anthropic's investigation suggests the attack was state-sponsored, emphasizing the need for enhanced cybersecurity measures as AI technology becomes more prevalent in such operations.

Health officials in the UK are warning of a severe flu season driven by a new H3N2 strain with mutations that could evade immune responses, leading to increased severe illnesses. The UK has seen an early start to the flu season, with H3N2 accounting for over 90% of cases analyzed. Jim Mackey, chief executive of NHS England, anticipates significant challenges for health care staff, while Antonia Ho from the University of Glasgow highlights the strain's potential for more severe illness, particularly in older adults. The US is also at risk as the flu season progresses.

Minnesota is exploring wastewater sampling as an early warning system for measles, following a successful test in Rochester where traces of the virus were detected during a small outbreak. The Minnesota Department of Health plans to publicly report measles levels in wastewater, building on techniques developed during the COVID-19 pandemic. This approach could help identify hidden cases, especially as vaccination rates decline, potentially leading to more severe measles cases.