Expedia is integrating ChatGPT into its app to help users make travel plans
Travel booking giant Expedia is integrating ChatGPT into its app as a beta test starting Tuesday to see how users interact with the ability to use generative AI when it comes to planning trips.
The chatbot, available in English on iOS, can give advice on where to go, where to stay, what to see, and what to do. The tool will also save any hotels discussed during the conversation in the app to remind bookers what was addressed.
“We view this as a really interesting capability to make the shopping experience, the discovery experience, easier for certain people who want to use natural language,” says Expedia CEO Peter Kern. “It may not be for everybody, it may not be for all of our consumers. We don’t know yet.”
Last month the company launched a plugin for ChatGPT, allowing consumers to use Expedia as they interacted with the language model. Now it’s bringing the capabilities into its own app. Expedia hasn’t invested in OpenAI, the maker of ChatGPT, but Kern said the two companies are both on separate “learning journeys” to figure out how generative AI can fit into travel and search more broadly.
“Generative AI definitely opens up a whole new spectrum of opportunities and possibilities for us to iterate and build on AI and our models that can truly transform travel over the next five years,” Expedia CTO Rathi Murthy tells Fast Company.
Over time, Expedia expects to bring the in-app function to Android and other languages. It won’t operate in countries where ChatGPT isn’t yet launched (or is banned, like Italy), Murthy adds.
The function comes at a time when governments and consumers are debating whether ChatGPT and other natural language models need to pause development in order for society to get a grasp on the technology and create safety guidelines.
Murthy says Expedia has been working with AI for quite some time and built up what she believes is the largest amount of data in the travel industry. The company has already put together a data council and architecture guild to examine each system it has in place for how it produces or consumes data.
“We’ve made sure we have transparency into how AI-powered decisions are made, and how predictions are arrived at . . . making sure we have the right governance in place to eliminate all sorts of bias from our AI outwards,” Murthy says. “This has been a practice that we have really matured over the years.”
(12)
