This week’s list of top data news highlights covers August 9, 2025 to August 15, 2025 and includes articles on preventing heat illness with wearable patches and detecting recycling contamination.
Tahoe Therapeutics, a biotech company based in California, has released Tahoe-100M, a dataset of one hundred million datapoints detailing how various cancer cells respond to more than 1,000 molecules. The data, generated using the company’s Mosaic platform, combines single-cell profiles from multiple patients and organs in each experiment, enabling large-scale mapping crucial for training AI drug discovery models. The dataset has already helped power an open-source virtual cell model, which achieved twice the accuracy of comparable AI systems.
Stanford University researchers have developed a brain implant capable of decoding and vocalizing words imagined by people with severe paralysis. Using electrode arrays in the motor cortex, AI models were trained to recognize neural patterns for phonemes during both attempted and imagined speech, achieving up to 74 percent real-time accuracy. The system includes a password lock to protect private thoughts, demonstrating potential for faster and more natural communication for people unable to speak.
Amazon is using a hybrid form of artificial intelligence called neurosymbolic AI, which combines neural networks that recognize patterns with symbolic reasoning that applies logical rules. Neural networks handle tasks like interpreting images or text, while the reasoning system checks the answers for accuracy, reducing errors by up to 99 percent. This setup powers Amazon’s Vulcan warehouse robots, which figure out where to pick up and place items, and its Rufus shopping assistant, which improves product recommendations by verifying them against rules about what is correct or relevant.
Emory University researchers have developed a wearable biopatch that attaches to a farmworker’s chest and monitors vital signs such as skin temperature, heart rate, and respiratory rate. The device is intended to detect early signs of heat-related illness and prevent potentially life-threatening conditions among agricultural workers. The system offers a low-burden, real-time monitoring solution designed to improve worker safety in high-heat environments.
5. Developing AI-Enabled Prosthetics
Cure Bionics, a medtech startup in Tunisia, has developed lightweight, muscle‑controlled prosthetic limbs using 3D printing and electromyography sensors. These devices interpret muscle signals to enable intuitive movement, offering a non‑invasive and more accessible alternative to traditional prosthetics. The technology aims to bring improved mobility and affordability to amputees, particularly in regions with limited access to advanced prosthetic care.
6. Producing Drinking Water from Air
AirJoule Technologies, a cooling and water technology company based in Montana, is deploying its first municipal system in Hubbard, Texas, to generate distilled water from air using heat recovered from a geothermal well. The system relies on a proprietary metal-organic framework to capture and condense water vapor, producing chemical-free water without refrigerants and with lower energy use than conventional methods. Once certified to meet drinking water standards, the project will supplement Hubbard’s municipal supply and demonstrate potential for pairing with waste heat sources such as data centers and manufacturing plants.
7. Detecting Recycling Contamination
Prairie Robotics, a sustainable recycling company based in Canada, has deployed AI-enabled camera systems on recycling collection trucks to identify contamination in curbside materials in real time. The system pinpoints the source address of contamination, then triggers personalized education such as postcards, app notifications, or in-person audits, helping municipalities reduce recycling errors without fines. Operating in about 40 North American cities, the technology also provides haulers with detailed route-level data to improve recycling quality and lower contamination-related costs.
Lawrence Livermore National Laboratory in California has developed a real-time tsunami prediction system using El Capitan, a supercomputer capable of 1.74 quintillion calculations per second. The system creates a digital twin of the ocean, transforming data from seafloor sensors into rapid shoreline impact forecasts using partial differential equation models. The approach delivers predictions in seconds instead of days, offering faster warnings to guide emergency response and coastal policy planning.
9. Speeding Up Police Report Writing
Axon, a technology company based in Arizona, has launched Draft One, an AI-powered tool that generates first drafts of police reports from body camera transcripts, reducing report-writing time by up to 70 percent. Officers review and edit the AI-generated text before submission, with built-in prompts to encourage accuracy, while the system uses a fine-tuned version of ChatGPT to minimize factual errors. The technology aims to ease staffing shortages and free up officer time, though it has prompted legal and civil rights concerns over potential biases, errors, and transparency in law enforcement records.
Researchers at the Massachusetts Institute of Technology have used generative AI to design two new compounds that kill drug-resistant gonorrhoea and staph infections in laboratory and animal tests. Trained on the chemical structures and antibacterial effects of known compounds, the system screens 36 million molecules, filtering for novelty and safety, and assembles promising candidates atom by atom.
