In the UK, over a million people suffer from frailty or dementia, severely compromising their ability to travel in urban environments. This paper presents SafeStep, an AI-driven travel system designed to assist elderly users with their journeys. At the core of SafeStep is a novel travel graph representation that integrates route planning with predictive modeling. The system performs the following operations at each stage of a journey:
- Generates personalized failure scenarios using a combination of LLMs and the Anticip8 behavioral prediction engine;
- Proposes targeted interventions;
- Estimates the impact of interventions on outcome probabilities.
This process enables SafeStep to select interventions that maximize the likelihood of the user reaching their destination.
SafeStep was evaluated through experiments on travel graph generation and a field study involving 26 real-world journeys. Results showed that combining Anticip8 for failure prediction with GPT-based models for intervention evaluation yields the most reliable performance. User feedback indicated that SafeStep improves confidence and perceived safety during travel, although interface usability needs to be improved for the target demographic. In the future, we aim to improve and release SafeStep. The AI system developed for SafeStep could also be applied in other areas such as mental health, career coaching, and addiction treatment.
Blogger's Review: SafeStep demonstrates the immense potential of AI in enhancing the travel experience for the elderly. By providing personalized failure predictions and interventions, it significantly boosts users' confidence and sense of safety. However, the usability of the interface still needs optimization to truly meet the needs of older adults. The future applications of this technology are promising and worth looking forward to.