This week’s list of top data news highlights covers May 10, 2025 to May 16, 2025 and includes articles on using AI to scout youth soccer players and robots that identify packages through touch.
Stanford University researchers have found a faster way to design enzymes by using AI to pinpoint the most effective DNA mutations. Traditionally, scientists engineer DNA sequences, insert them into cells, grow those cells to produce the enzymes, and then test the enzymes for effectiveness—a process that takes months. Instead, the Stanford team trained AI models on data from thousands of past enzyme experiments to predict which specific DNA edits would boost activity. The models output a shortlist of precise DNA edits, which researchers synthesize as short DNA templates. Those templates go into a simple test-tube reaction to build and assay each enzyme variant. This pipeline takes days instead of months and has produced enzymes up to 40 times more active than the originals.
2. Combatting Healthcare Data Gaps
OMNY Health, a health data platform based in Georgia, has compiled a dataset of over 85 million deidentified patient records and 4 billion clinical notes across 200 specialties. The company makes this data available through a secure platform that allows users to search, filter, and analyze real-world clinical information. Researchers and health tech firms are using the data to train AI models to predict disease risk, prevent hospital readmissions, and improve representation in clinical trials.
3. Digitizing Soccer Recruiting
Footbao, a Brazil-based sports tech startup, has developed an app that uses AI to help clubs scout youth soccer talent more efficiently. Players upload videos of themselves completing drills or playing in matches. The AI model compares their performance to a growing database of top-tier athletes, identifying key skills and areas for improvement. Footbao’s in-house analysts then review the most promising profiles and recommend players to partner clubs based on their specific needs.The platform is designed to widen access to professional opportunities by making talent discovery faster, more precise, and less dependent on geography or income.
Helsing, a defense tech company based in Germany, has developed an autonomous underwater drone designed for long-term naval surveillance. The drone can patrol for up to three months and its AI system can detect acoustic signatures of ships 40 times faster than human analysts and operates at sound levels 10 times quieter than prior models. Designed to counter growing threats to European subsea infrastructure, the drone is slated for deployment within a year.
Researchers at Mass General Brigham have developed FaceAge, an AI tool that estimates a patient’s biological age from a facial photo. Rather than focusing on chronological age, the model assesses signs of physical aging to give doctors a clearer picture of a patient’s underlying health. In a study of cancer patients receiving palliative radiotherapy, doctors used FaceAge scores to predict who was likely to survive the next six months. With the AI system’s input, their prediction accuracy jumped from 61 percent to 80 percent.
General practitioners at the Jean Bishop Centre in England are using an AI tool that transcribes consultations, drafts referral letters, and generates care summaries in real-time. The system reduces administrative workload by at least half for practitioners, freeing up time for more direct patient care. Staff using the tool report that it frees up their time to stay current with medical learning and helps prevent burnout.
Researchers at the University of Oklahoma, in partnership with startup Xironetic, have developed an augmented reality tool that helps neurosurgeons visualize a patient’s vascular system during surgery. Using a headset, surgeons see a 3D hologram aligned with the patient’s anatomy, showing blood vessels that are normally hidden from view. This is especially useful in procedures like external ventricular drain (EVD) placement, which are often performed without direct visual guidance. The tool aims to reduce complications, improve precision in high-risk cases, and enhance training for early-career surgeons.
8. Sensing Packages with Robotic Arm
Amazon has introduced Vulcan, a warehouse robot equipped with tactile sensors along its joints to help it identify and retrieve objects by touch. The sensors feed data to a machine learning algorithm that interprets object shape and texture, helping the robot navigate shelves and pick items accurately. Already deployed in Germany and Washington state, Vulcan aims to reduce the physical strain on human workers and take on one of the more difficult warehouse tasks—locating and picking individual products from densely packed bins.
9. Understanding Urban Dynamics
Researchers at Manchester Metropolitan University in the UK have developed an AI tool that uses geotagged social media posts to map and forecast real-time urban behavior. The system collects public posts tagged with time and location, then uses machine learning to group them into activity clusters and analyze their content. By identifying where and how topics like work, nightlife, or public gatherings change over time, the tool reveals shifting patterns in how cities respond to disruptions. Tested during the COVID-19 lockdowns, it showed how people adapted their routines.
10. Observing Pollution Concentration
Researchers at Istanbul Technical University have created an AI model that estimates vehicle emissions and traffic-related air pollution using surveillance camera footage. The system identifies vehicle type, speed, brand, and model, then calculates emissions using standardized emission factors. These estimates are combined with weather and airflow data in a fluid dynamics model to simulate how pollutants disperse across urban areas, offering city planners real-time insight into air quality by location and time of day.
