This week’s top data news highlights from August 2, 2025 to August 8, 2025 include advances in extending lithium battery life and AI-driven reconstruction of fragmented Latin texts from Ancient Rome.
1. Merging Virtual and Real Touch
Graduate students at Carnegie Mellon University have created Luceal, a tactile AI-powered pet that exists in both VR and the physical world. The system combines Apple Vision Pro headsets with custom textile sensors in a plush animal that detect touch and send signals to trigger lifelike animations and response in VR. By integrating soft sensor data with AI-driven animation, Luceal reacts in real time to user interactions. This fusion of tactile and virtual engagement offers a new way to experience companionship without the responsibilities of traditional pet ownership.
Tapology, U.S.-based combat sports platform, has launched a data-driven ranking system that evaluates UFC fighters using in-cage performance data, rather than relying on media opinion or fan voting. An algorithm analyzes each fighter’s last six UFC bouts, factoring in things like short-notice matchups and the strength of competition to generate unbiased rankings within each weight class. By removing human subjectivity, Tapology’s system offers a more transparent view of which fighters are earning their way toward a title shot.
3. Expanding Lithium Battery Life
Researchers at the University of California, San Diego have developed an algorithm that brings consistency to how scientists evaluate lithium battery health. Traditionally, researchers use scanning electron microscope images to visually judge how lithium spreads across a battery surface, but these images are interpreted differently from lab to lab. The new algorithm analyzes the images by converting them into black-and-white pixel maps, quantifies how lithium is distributed, and calculates a score that shows how evenly lithium is spread. This makes it easier to track battery health and compare results across studies.
Google’s Gemini app now includes Storybook, which lets users generate custom illustrated stories using prompts, uploaded images, and narration in over 45 languages. The system uses language and image generation models to create coherent, visually rich 10-page narratives. Users can choose art styles like claymation or comic book formats. It’s a powerful demonstration of how multimodal AI can turn personal ideas into tailored storytelling experiences.
5. Designing Tougher Plastics with AI
MIT researchers trained an AI model to predict how a polymer’s molecular structure affects its flexibility, strength, and durability. The system simulates thousands of polymer combinations before any physical testing begins. It narrows down the best candidates with the desired properties. This speeds up materials discovery while reducing development costs and environmental waste.
6. Controlling an iPad Through Thought
Synchron, a brain-computer interface company, has demonstrated how a patient can control a mobile device using their thoughts using an implant placed in a blood vessel near the brain’s motor cortex. The device records brain signals that reveal a person’s intention to move, and then the technology uses AI to translate these signals into touchscreen commands using Apple’s brain-computer interface protocol. The system provides real-time feedback to align mental intent with digital action. This breakthrough could dramatically expand device accessibility for people with severe motor impairments.
7. Merging Data Artistry with Storytelling
Digital media design studio Exhibitry has created a new immersive LED installation at Purdue University that processes over 2.8 million pixels to create perspective-based, 3D-like visuals tied to key engineering milestones. It uses viewer-position tracking and a flexible content management system to adapt animations in real time. Content includes events like the moon landing and AI-powered race cars. The display offers a striking example of how data-rich media can enhance education and public engagement.
8. Cutting Vehicle Emissions at Intersections
MIT researchers created an AI model to simulate one million real-world driving scenarios and optimize vehicle speeds when approaching intersections. Their model recommends eco-driving strategies that smooth acceleration and braking to reduce stop-and-go behavior. Full adoption could cut carbon dioxide emissions at intersections by up to 22 percent. Even limited use could yield significant emission reductions without slowing traffic.
Researchers at the University of Southern California developed the MOTIF Hand, a robotic platform equipped with multimodal sensors to detect heat, pressure, and weight with human-like precision. Using a thermal camera in the palm, fingertip force sensors, and motion-based weight estimation, the hand evaluates an object’s properties before making contact. In tests, it safely handled hot items, distinguished between objects of equal size but different weights, and adjusted its grip accordingly. This innovation pushes robotics closer to human-level dexterity and touch sensitivity.
Google DeepMind trained its Aeneas model on over 176,000 Latin inscriptions to predict missing content and estimate the likely origin and age of incomplete texts from ancient Rome. The model analyzed both the text and image data to find linguistic and stylistic matches across time and geography. Historians say it could dramatically accelerate inscription analysis and open up new avenues for classical research.
Image credit: Pixabay
