An interdisciplinary UMD team will investigate why AI falls into human-like reasoning traps—and how to make it more reliable.

The five-year award expected to total $37.5M supports the institute’s next phase of research, education and workforce development.

The DOE-funded project will combine generative AI and Earth system modeling to deliver faster, more detailed forecasts for floods, droughts and water resources.

Researchers show how targeting and performance data can combine to reveal sensitive information.

UMD’s Connor Baumler finds that editing AI-generated text may not erase its stylistic fingerprints.

QuICS Fellow Nicole Yunger Halpern collaborates on research identifying which quantum states can deliver a true computational advantage.