From lab data to underwater glue: how AI sped up Japan’s hydrogel breakthrough

Turning raw chemical data into a working material usually takes years of trial and error, but researchers at Osaka University and collaborating institutions in Japan say an AI-driven framework helped them do it far faster — producing a hydrogel with record-setting underwater adhesion and self-healing ability.

The team’s approach combined machine learning, data mining and high-throughput laboratory experiments, letting an AI model analyse large datasets to predict which molecular structures would perform best before those predictions were synthesised and tested experimentally. The study, published in Nature as ‘Data-driven de novo design of super-adhesive hydrogels,’ describes the result as “an AI-driven materials discovery framework for multifunctional hydrogels.”

The resulting hydrogel achieved underwater adhesive strength exceeding 1 MPa, among the strongest reported in its class, while retaining high mechanical toughness, elasticity and the ability to repeatedly heal itself after damage — properties that conventional hydrogels have historically struggled to combine, especially underwater.

Researchers believe the material could find use in medical adhesives, wearable electronics, soft robotics and underwater repair technology, calling it an example of how AI-assisted discovery is reshaping the pace of materials science research.