Guide dogs have long helped people with vision loss move safely through the world. They can avoid obstacles, follow routes, and respond to commands.
But one thing they cannot do is talk—until now.
Researchers have developed a new kind of guide dog that not only leads users but also speaks to them, offering guidance and explanations along the way.
A team at Binghamton University, part of the State University of New York, has created an AI-powered robotic guide dog that can communicate using natural language.
The system uses advanced artificial intelligence, including large language models, to understand voice commands and provide real-time spoken feedback.
The research was presented at the AAAI Conference on Artificial Intelligence in 2026.
According to project leader Shiqi Zhang, this technology goes beyond what real guide dogs can do.
While trained dogs can understand a limited number of commands, the robotic version can process complex instructions and have simple conversations with users. This opens up new possibilities for communication and independence.
The robot is designed to guide users to their destination while keeping them informed about their surroundings.
Before starting a journey, it can describe possible routes and estimate how long each one will take.
This helps users choose the path that best suits their needs. Once the journey begins, the robot continues to talk, describing the environment in real time.
For example, it might say that the user is entering a long hallway or approaching an obstacle.
This feature is especially helpful for people who are blind or have low vision, as it provides a clearer sense of what is happening around them.
Instead of simply following directions, users can better understand their environment and feel more in control.
To test the system, the researchers worked with seven participants who were legally blind. They asked the robot to guide them through a large office space to a specific destination. During the test, the robot explained the available routes, led the participants safely, and described the surroundings along the way.
After completing the task, participants shared their feedback. Most of them preferred a combination of route explanations before the journey and ongoing descriptions during the walk. They found this approach helpful, easy to understand, and more informative than traditional guidance alone.
The results suggest that adding speech to robotic guide systems can significantly improve user experience. It not only helps with navigation but also increases confidence and awareness.
Looking ahead, the research team plans to improve the system further. They aim to make the robot more independent, able to travel longer distances, and capable of working in outdoor environments as well as indoors.
The participants were enthusiastic about the technology and eager to see it become part of everyday life.
If developed further, talking robotic guide dogs could offer a new level of support for people with vision loss, combining the reliability of machines with the clarity of human-like communication.
