Uber integrates OpenAI to power AI assistant and voice features for drivers and riders

Uber has deployed OpenAI's large language models across its platform to create AI-powered tools that help drivers optimize earnings and riders book trips more efficiently. The company developed Uber Assistant, a multi-agent AI system that provides real-time marketplace insights—such as where to position themselves or whether to switch between rides and deliveries—by analyzing complex data like traffic, weather, and local events. Built on three core principles of safety, trust, and low latency, the system uses a governance layer called AI Guard to screen prompts and responses. Uber also introduced voice booking via OpenAI's Realtime API, allowing users to request rides using natural speech, which is especially beneficial for travelers with complex needs, older adults, or visually impaired riders. The voice feature is rolling out over the coming weeks. Uber Assistant is currently in experimental rollout across the U.S. driver network, with hundreds of thousands of drivers accessing beta experiences. Early results show accelerated ramp-up for new drivers and strong repeat engagement from experienced drivers. Uber's engineering teams now embed AI capabilities across the organization rather than relying on a centralized team, enabling faster iteration. The collaboration underscores how LLMs can address real-time marketplace challenges at scale, processing 40 million trips daily across 15,000 cities in over 70 countries.