This week’s list of top data news highlights covers March 22, 2025 to March 28, 2025, and includes articles on searching for mineral deposits using AI and changing accents in real-time.
1. Improving Weather Forecasting
Academics at the Alan Turing Institute in the UK have developed Aardvark, a weather model that offers a faster and more efficient alternative to traditional forecasting methods. Conventional forecasts rely on supercomputers to simulate atmospheric physics based on observational data, a process that is computationally intensive and time-consuming. Aardvark takes a different approach: it uses AI to learn patterns from historical weather data and generate accurate forecasts directly, without simulating the physics. As a result, it can produce forecasts in seconds using far less computing power.
2. Fast-Tracking Drug Discovery
Researchers at the Oxford Drug Discovery Institute in the UK are using AI-powered databases to accelerate drug discovery for Alzheimer’s.These tools help scientists quickly scan and make sense of massive amounts of scientific literature and data, highlighting connections that might otherwise be missed. The team used this approach to narrow down 54 genes related to Alzheimer’s as likely targets for lab testing.
3. Simulating Naval Operations
The U.S. Navy is touring schools with its Strike Group Mixed Reality Experience, an interactive program that helps students gain a better understanding of Navy careers. By using VR technology, students can participate in simulations from driving coastal craft to welding underwater, offering a hands-on look at life in the Navy before deciding whether to enlist.
4. Identifying Untapped Mineral Deposits
Earth AI, a predictive mining company based in California, is using AI algorithms to identify overlooked deposits of critical minerals. The algorithm is trained on 400 million geological reports and exploration records, allowing it to scan wide areas of land and predict where valuable metals like copper, cobalt, gold, and silver may be. The company is currently focusing on underexplored regions in Australia, where it has already identified promising sites in the Northern Territory and New South Wales.
Veesion, a Paris-based AI company, has developed surveillance software that uses machine learning and motion analysis to help retailers catch shoplifters. The system scans closed-circuit TV footage for suspicious body movements, such as hiding items or opening packaging, using biometric signals like posture and gestures. Now in use at more than 4,000 stores across 25 countries, the technology has helped cut shoplifting by around 60 percent.
6. Diagnosing Cancer Accurately
Researchers at Leeds University in England have created a small robot that can be used to detect and potentially treat cancers found in the large intestine. Unlike traditional screenings such as colonoscopies, which require inserting a long tube and sometimes taking tissue samples, the robot moves through the digestive system guided by a magnet outside the body and creates detailed 3D scans. The robot could streamline diagnosis and treatment into a single, less invasive process.
NASA’s Parker Solar Probe is returning to Earth after collecting information about the Sun from the closest point a probe has ever been. The probe uses advanced sensors and instruments to collect sample particles and record information about the Sun’s magnetic field. The data collected by the probe can be used to provide information on solar wind, the Sun’s structure, and information on how clouds of ionized gas that can cause disturbances of Earth’s magnetic field are constructed.
8. Changing Accents in Real-Time
Krisp, an AI company based in Armenia, is using AI to change user accents during phone calls in real-time. The model was trained on thousands of speech samples featuring different accents and dialects, allowing it to preserve the speaker’s voice while adjusting phonemes, the distinct set of sounds that distinguish different words, to create an American accent. When tested with enterprises, sale conversion rates increased by roughly 26 percent and revenue per booking increased by 14 percent.
9. Automating Recruitment Workflow
SourceWhale, a UK-based software company, has built an AI tool that helps recruitment agencies automate outreach and follow-ups. The platform connects to systems recruiters already use, like LinkedIn, email, and applicant tracking systems, and pulls information from them to automatically send messages, track responses, and update records. By removing manual admin, it helps recruiters act faster and more efficiently.
10. Streamlining Coeliac Diagnosis
Scientists at the University of Cambridge in the UK have developed an AI model that speeds up the diagnosis of coeliac disease, an autoimmune condition triggered by gluten. The algorithm was trained on 4,000 biopsy images of the small intestine from 5 different hospitals. It matches the accuracy of human pathologists but delivers results in under a minute, helping reduce diagnostic delays.
