Airbnb Halts ChatGPT Integration, Relies on Qwen for AI Support
Airbnb Pauses ChatGPT Integration Amid Platform Concerns
Airbnb has temporarily halted plans to integrate OpenAI's ChatGPT into its platform, citing insufficient maturity in the AI tool's connection capabilities. CEO Brian Chesky revealed this decision in a recent interview, emphasizing the need for stronger platform support before proceeding.

Future Collaboration Remains Possible
While postponing immediate integration, Chesky indicated Airbnb continues monitoring ChatGPT's ecosystem development. The company may follow models similar to Booking Holdings or Expedia Group when technical conditions improve. Notably, Chesky maintains a close relationship with OpenAI CEO Sam Altman, having previously advised on ChatGPT's third-party developer features.
Internal AI Advancements Accelerate
Simultaneously, Airbnb is enhancing its proprietary AI systems:
- Fully upgraded AI customer service launched nationwide in May
- Processes order modifications/cancellations without human intervention
- Reduced human support demand by 15%
- Slashed average response time from 3 hours to 6 seconds
The system utilizes 13 different AI models, including solutions from:
- OpenAI
- Alibaba (Qwen model) - primary workhorse
- Various open-source providers
"We rely largely on Alibaba's Qwen model," Chesky noted. "It performs exceptionally while maintaining cost efficiency. We use OpenAI's models more selectively where alternatives aren't as effective."
Strategic Approach to AI Adoption
Industry analysts observe Airbnb pursuing a multi-model strategy prioritizing:
- Operational flexibility
- Cost optimization
- Performance benchmarking
The company appears unwilling to become dependent on any single vendor despite existing relationships with major AI providers.
Key Points:
- ❌ ChatGPT integration delayed over platform immaturity concerns
- 🤖 Internal AI systems already handling 15% more queries autonomously
- 💰 Cost-effective Qwen model forms backbone of customer service AI
- 🔮 Future OpenAI collaboration remains possible pending technical improvements

