USC computer scientists led by Yue Zhao, assistant professor at USC Viterbi School of Engineering and the USC Mark and Mary Stevens School of Computing and AI, have developed tools to audit and monito…
A study of more than 15 large language models by Dr. Filippo Menczer of Indiana University Bloomington found that AI fact-checking tools are less accurate than human fact-checkers but can achieve 98-9…
A USC-led study found that advanced audio large language models (LLMs) such as those in ChatGPT and Gemini struggle to interpret nonverbal cues like tone and emotion, often taking speech literally and…
A new University of Southern California research study, funded by the National Science Foundation, will monitor brain activity of doctors, journalists, and software engineers using electroencephalogra…
USC researchers developed an AI tool to identify tweets containing COVID-19 misinformation and distinguish users spreading false claims from those countering them, finding that counter-misinformation …
USC researchers will present three papers at the 2026 Robotics: Science and Systems (RSS) conference in Sydney, Australia, covering robotic foundation models, humanoid loco-manipulation, and robot man…
A USC-led study used AI to analyze biomarkers including sweat, brain signals and eye movements to classify depression and suicidal ideation, offering a potential objective testing method for mental he…
A USC study evaluated AI chatbots ChatGPT-4, Llama 3.3, and Gemini 1.5 Pro on mental health questions, finding that while AI offers accessible support, licensed professionals rated responses as often …
A former Meta creative director who now lectures at colleges across the country says the Class of 2026 is the most technically gifted graduating class in a generation, yet graduates cannot find jobs. …
Self-represented litigants are using AI chatbots to flood U.S. federal courts with professionally formatted but often meritless lawsuits, causing docket entries to spike 158% above pre-AI averages. MI…
A study by MIT researcher Anand Shah and USC researcher Joshua Levy found that 18% of pro se court filings now contain AI-generated text, with the volume of pro se docket entries per court in the firs…
The article describes a two-layer approach to prevent AI assistants from generating false information, using NVIDIA NIM. The first layer tightens the system prompt to define what the assistant is allo…
The article describes how to upgrade a simple AI campus assistant from using a hardcoded knowledge base in the prompt to a proper Retrieval-Augmented Generation (RAG) system using NVIDIA's hosted embe…