New AI models are surprisingly good at 'reading' human minds

(Web Desk) – Researchers at Texas A&M University's College of Engineering and Korea Institute of Science and Technology have introduced a new artificial intelligence (AI) system called OmniPredict designed to improve the safety of self-driving cars.

OmniPredict is the first system to predict pedestrian behavior using multimodal large-scale language models (MLLM). It leverages the same kind of underlying technology used in advanced chatbots and image recognition, but its goals are different.

The system aims to predict in real time what a person is likely to do next by combining what they see and the details of the situation.

Early tests have attracted attention, showing that OmniPredict can achieve significantly higher accuracy without special training.

“Cities are unpredictable. Pedestrians are unpredictable,” said Dr. Shinkanth Saripalli, principal investigator on the project and director of the Center for Autonomous Vehicles and Sensor Systems. “Our new model offers a glimpse into a future where machines not only see what's happening, but also predict what humans will do.”

As developers push to improve the safety of autonomous driving, OmniPredict adds a new layer of road awareness that approaches human-like intuition.

It not only reacts to the pedestrian's current movements, but also tries to predict what that person will do next. If it works as intended, this approach could impact how self-driving cars operate in dense urban environments and navigate busy roads more smoothly.

“This opens the door to safer autonomous vehicle operation, fewer pedestrian-related accidents, and a shift from reactive to proactive hazards,” Saripari said.

“We are opening the door to exciting applications,” Saripari said. “For example, the potential for machines to detect, recognize, and predict the consequences of humans exhibiting threatening signs could have important implications.”