Research News

RIT is hosting four Research Experience for Undergraduate (REU) program sites this summer, which are designed for students to develop a passion for research and receive guidance for pursuing graduate careers. The university has a longstanding history hosting REU sites, which are funded by grants from the National Science Foundation.

A team of RIT researchers, led by Professor Linwei Wang, is improving healthcare by creating hybrid AI digital twins of patient’s organs and diseases. The work will lead to real-time, physics-integrated digital twins of the liver and improved AI models that could assist doctors with heart surgery.

Like Indiana Jones, researchers at RIT’s Interaction, Media, and Learning Lab hope to save pieces of history before they disappear so that people can learn from the designs.

RIT faculty members are tackling some of AI’s most pressing challenges with support from the National Science Foundation’s prestigious CAREER Award program, which recognizes early-career scholars who integrate cutting-edge research with student education.

RIT computer science professor Weijie Zhao has earned a National Science Foundation CAREER Award to defend machine learning from adversarial threats.

The University of Mississippi writes about a new book from Andrea Hickerson, dean at Ole Miss and former RIT faculty member; Christopher Schwartz, research scientist in the Department of Cybersecurity; and Matthew Wright, O’Sullivan Professor and chair of the Department of Cybersecurity. The book, Fake-Checking: A Journalist's Guide to Deepfakes, is a practical guide that addresses growing concerns of AI-generated media's impact on accurate reporting.

Cybersecurity experts are studying what happens when autonomous AI systems collect, process, or share Social Security numbers and other sensitive data. The published research finds that many AI tools lack safeguards and don’t align with privacy policies.

The Decoder reports on research by RIT that examines how many human evaluators are needed for reliable AI benchmarks, finding that more than 10 raters per example are often required to capture meaningful differences in model performance.

RIT has launched a new decade-long strategic framework that will guide the university’s priorities, investments, inspiration, and aspirations through 2035.

Republic World reports on RIT’s HAUNT framework, which tests how large language models respond to misinformation.

Researchers have created a new tool to stress test large language models (LLMs) that power AI chatbots, including ChatGPT. Using this innovation, developers can build more trustworthy AI systems and users can see which LLMs contradict themselves.

As 2025 comes to a close, we’re reflecting on a year filled with milestones, innovation, and inspiring stories from across the RIT community.