Kan AI nieuwe duurzame materialen ontwikkelen ?

De ontwikkeling van nieuwe materialen is cruciaal voor het bevorderen van technologieën en het verminderen van onze ecologische voetafdruk. AI wordt op deze uitdaging toegepast, met de mogelijkheid om nieuwe materialen met unieke eigenschappen te ontdekken. Door enorme hoeveelheden data over materiaalsamenstelling en -eigenschappen te analyseren, kan AI het gedrag van nieuwe materialen voorspellen en combinaties suggereren die nog niet eerder zijn uitgeprobeerd. Dit zou kunnen leiden tot doorbraken op gebieden zoals energieopslag, bouw en elektronica. Het gebruik van AI in de materiaalkunde belooft ook het ontdekkingsproces te versnellen, waardoor de tijd en kosten die gepaard gaan met traditionele trial-and-error-methoden worden verminderd. Naarmate de wereld op zoek gaat naar duurzamere oplossingen, wordt de rol van AI bij materiaalontwikkeling steeds belangrijker.

Background

The development of new materials is crucial for advancing technologies and reducing our environmental footprint. AI is being applied to this challenge, with the potential to discover novel materials with unique properties. By analyzing vast amounts of data on material composition and properties, AI can predict the behavior of new materials and suggest combinations that have not been tried before. This could lead to breakthroughs in fields such as energy storage, construction, and electronics. The use of AI in material science also promises to accelerate the discovery process, reducing the time and cost associated with traditional trial-and-error methods. As the world seeks more sustainable solutions, the role of AI in material development is becoming increasingly important.

AI is already contributing to the discovery of new sustainable materials by accelerating simulations and screening vast chemical spaces, for example using generative models to propose candidate molecules and density-functional theory to evaluate stability and performance. Recent systems like GNoME, MatterGen and AlphaTensor have identified thousands of stable inorganic structures and even novel superconductors with reduced trial-and-error, while robotics-driven labs such as those at DeepMind and Carnegie Mellon are closing the loop by autonomously synthesizing and characterizing promising candidates. Although human expertise remains critical for setting objectives and interpreting results, AI is demonstrably able to propose viable new materials faster than traditional methods, cutting design-to-discovery timelines from years to months.

— Enriched May 12, 2026 · Source: DeepMind