Harnessing the power of generative AI, researchers at Tsinghua University have developed AIGP—a diffusion-based generative framework that enables instant translation of optical properties into fabrication-ready metasurfaces. By using transmission, phase, and polarization as “prompts,” AIGP directly maps optical properties to subwavelength, fabricable structures, generating high-fidelity metasurface designs in seconds. This breakthrough overcomes critical bottlenecks in photonic inverse design and paves the way for large-scale, AI-driven generative optical devices.

Natural biomass hydrogels are emerging as promising building blocks for intelligent sensors because they combine softness, water-rich structures, tunable networks, and abundant functional groups. A new review brings together recent progress in turning materials such as cellulose, chitosan, sodium alginate, gelatin, starch, hemicellulose, proteins, and lignin into responsive sensing platforms. By focusing on the correlation between crosslinking networks and sensor behavior, the review connects material design with practical performance, including sensitivity, response speed, durability, and environmental stability. The work offers a clearer roadmap for developing sustainable sensors for wearable electronics, healthcare monitoring, environmental sensing, and smart human–machine interfaces.

A comprehensive review published in Skin outlines the emergence of “Dermatology AI 2.0”, a fundamental transition from pattern recognition to cognitive and actionable intelligence. Built on four core pillars—causal inference, skin digital twins, predictive intervention, and distributed autonomous networks—this new paradigm enables AI to diagnose rare diseases 30% more accurately, predict disease flares with >90% accuracy, and deliver full-lifecycle skin health management. The review emphasizes that AI will not replace clinicians but will automate routine tasks, allowing physicians to focus on complex cases and patient care.

Researchers from Florida International University have discovered that certain visual-language AI models can be manipulated via subtly manipulated images. 

Researchers developed a data-driven method that combines a GA-BP neural network with a chaotic particle swarm optimization algorithm to predict and optimize screen-printing parameters for thick-film resistors. The approach predicts resistance with an R² of 0.991, recommends optimal process settings in about 2.38 seconds, and cuts resistance error by up to 14.86%, keeping deviations within 5%.

The recent suspension of Anthropic’s Mythos and Fable models brings the conflict between AI innovation and safety into public view: some favor innovation prohibitions while others are skeptical of safety constraints. A new paper in Law, Ethics and Technology by philosopher James Brusseau resolves the dilemma with “acceleration AI ethics,” a framework for answering AI risks with further innovation. Using the Telus GenAI customer support agent as a case study, the article argues that responsible AI does not have to choose between moving fast and advancing safely. Instead, the safety risks generated by innovation are overcome by the same innovative forces that created them.