Advertisement

China travel company Trip.com Group creates own AI model, Wendao, to push artificial intelligence as a tool for accurate voyage data

  • The Shanghai-based group – operator of Trip.com, Ctrip, Qunar and Skyscanner – has created a travel-oriented large language model called Wendao
  • This AI model is expected to become an essential tool for its customers to obtain accurate travel data and deal recommendations

Reading Time:2 minutes
Why you can trust SCMP
Trip.com Group’s large language model initiative reflects the broader Chinese tech industry’s determination to close the gap with the West in building ChatGPT-like services. Photo: Shutterstock
Ben Jiangin Beijing
Trip.com Group, China’s largest online travel services provider, has launched its own large language model (LLM) – the technology used to train chatbots like ChatGPT – that is expected to make artificial intelligence (AI) an essential tool for its customers to obtain accurate voyage data and deal recommendations.
Advertisement
The Shanghai-based company – operator of online travel agencies Trip.com, Ctrip, Qunar and Skyscanner – unveiled its LLM, Wendao – which roughly translates to “ask for your ways” in Chinese – at a corporate event in its home city on Tuesday.
With tourism continuing to be “an important industry”, Trip.com co-founder and executive chairman James Liang Jianzhang said at the event that the group is committed to embracing “technological development and to use advanced AI-driven services to improve the travel process and overall experience”.

LLMs are deep-learning AI algorithms that can recognise, summarise, translate, predict and generate content using very large data sets.

Trip.com’s travel-oriented Wendao LLM, which is still in the testing phase and only available to a small group of users, was trained with 20 billion quality travel-relevant data sets, bolstered by the company’s real-time data and its proprietary search algorithm, according to Liang.

Advertisement

“The opportunity for vertical LLMs is that AIGC (artificial intelligence-generated content) has met a tremendous challenge – their responses are not reliable,” he said.

Advertisement