We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·May 6·5 min read
Uber uses OpenAI to help people earn smarter and book faster
TL;DR
Uber's drivers and riders faced inefficiencies: drivers lacked real-time earning guidance, and riders experienced slow booking. Uber deployed OpenAI's language models to power AI assistants and voice features integrated into their platform.
✦ Why It Matters
Engineers can apply LLMs and voice interfaces to reduce friction in two-sided marketplaces and improve real-time decision-making.
Key Takeaways
- Integrate LLMs into marketplace platforms to provide real-time, personalized recommendations to users.
- Use voice interfaces powered by language models to reduce friction in user interactions and decision-making.
- Apply AI assistants to two-sided markets to optimize outcomes for both supply and demand sides simultaneously.
- Leverage natural language processing to contextualize user data and deliver actionable, timely guidance at scale.
Full Summary
Uber operates a real-time marketplace connecting millions of drivers and riders worldwide, but faced two key friction points: drivers made earning decisions with incomplete information, and riders experienced delays during booking. Uber built AI assistants powered by OpenAI's large language models (LLMs)—neural networks trained on vast text data to understand and generate human language—and integrated voice interaction features into their mobile applications.
The solution leverages OpenAI's models to process driver context (location, demand patterns, earnings data) and rider requests (destination, preferences, timing) in natural language, enabling personalized recommendations. Drivers receive real-time suggestions on where and when to drive for optimal earnings, while riders benefit from faster, more intuitive booking flows through voice commands and conversational interfaces.
This deployment demonstrates how LLMs can optimize two-sided marketplace operations by reducing decision friction and improving user experience at scale.