It could also be used to develop strong, lightweight components for airplanes or modular robots, or architectural elements like sound-dampening panels and building facades.
J-PAL North America launches a free resource hub to provide evidence-informed tools for reducing homelessness and improving housing stability. The hub aims to equip decision-makers with foundational resources on leveraging evidence to reduce homelessness and improve housing stability. It is a public repository of multimedia content developed with a grant from Google.org. The hub is hosted by J-PAL North America on Oct. 27 and aims to increase access to evidence-informed paths to housing stability.
Google Research has developed open foundation models for tuberculosis screening, moving from early research benchmarks to open foundation models that can be deployed at the point of care. This week at IDWeek 2026, Google is sharing how these models are empowering public health systems worldwide to democratize TB screening at global scale.
The debate, then, was about trying to determine what human inventiveness boils down to, and then how to reconstitute human inventiveness in machine form.
Jiaji Li and his team at TJU and CSAIL used X-Hinges to create interactive objects like a glove that teaches a robot how to grasp blocks and a lamp that adjusts brightness based on user interaction. They also developed a device that identifies objects placed on it based on tactile data. The interface allows users to design interactive objects from scratch, with four shapes to choose from and customizable movements. The advanced physical awareness could be integrated into detection systems in various settings.
Researchers at MIT, Boston Children’s Hospital, and Inria have developed a deep learning model called E(3)-Pose to help image the fetus's head during prenatal ultrasound scans. The model aims to capture high-resolution images of the fetus's head, reducing the risk of missed or incorrect diagnoses of developmental abnormalities, particularly in the brain. The work was presented at the 2026 European Conference on Computer Vision.
Google Research is developing AI agents to navigate complex biological landscapes in wet labs, aiming to discover single-digit nanomolar protein binders for cancer targets. The initiative addresses two challenges: reliable reasoning about 3D geometry and physical state in specialized robots, and real-world feedback for hypothesis validation in wet-lab experiments. The projects highlight progress in bridging AI exploration and experimental validation.
Researchers at the National Institute of Standards and Technology (NIST) and their colleagues have demonstrated a new way to generate a miniature tape measure for light.
Governments can rebuild public trust by partnering with trusted private organizations, as shown by a study in Kenya. The study found that highlighting the role of a telecommunications company increased portal usage but hurt trust in government. This approach can help solve problems requiring public cooperation and improve service delivery efficiency.
Google Research is showcasing tools that put never-before-possible insights and capabilities into the hands of scientists, developers, and others. The company is expanding access to Google Earth AI, a planetary intelligence tool, and introducing new experimental features for Google Earth AI.
Deep Review, an AI system designed for rigorous verification of complex documents, including Google Docs and presentations, has launched publicly. Initially designed for academic papers, its capabilities have expanded to cover various domains, including legal and finance. It can be used as an iterative improvement tool during drafting and a verification tool during review. The system aims to address the validation bottleneck in content creation, where the human capacity for verification has remained fixed amid rapid AI-generated content.