In the past almost two years, tech giants like Google, Meta, and OpenAI have showcased how generative artificial intelligence, or gen AI, can potentially help users plan their travel itineraries. Google recently announced the development of trip-planning features and tools to discover reviews and destination details.
Most established online travel platforms, like Expedia, Booking, Kayak (with their custom GPT add-ons), and Skyscanner, are also exploring how gen AI can enhance their existing travel planning services. The current implementations focus on chatbots to answer users’ questions about destination itineraries—like planning a three-day trip to Paris in July.
Despite the optimism and general interest, there are concerns about whether the chatbot interface provides the right travel planning experience and whether the AI models’ information is reliable.
The primary issue is that generative artificial intelligence can make substantial errors in its responses or “hallucinate,” i.e., invent incorrect information or data. This happens because these models learn to make predictions by finding patterns. However, incomplete or biased training data can lead to prediction mistakes or hallucinations.
Discover how generative AI is transforming travel planning with insights from Rome2Rio’s CEO.
All major tech companies acknowledge this problem and caution users to verify the information they generate.
I recently heard an illustrative story from an Australian tech executive who used one of the leading gen AI platforms to plan a trip to Japan. The model suggested visiting three cities. While two of these destinations turned out to be lovely, the AI invented the third one, which was the most intriguing and appealing suggestion.
The tool was so persuasive that the executive was quite disappointed when he discovered the hallucination.
This scenario contrasts with how current digital travel providers operate. At Rome2Rio, for example, we collect and verify transport and schedule information from over 12,000 companies across more than 240 countries and territories, including airlines, trains, buses, ferries, taxis, and rideshare operators.
To gather and check the data, we combine APIs and digital feeds with human curation and verification from a network of global collaborators.
Accurate, timely information is the bedrock of our business, as our users rely on us to make decisions. Most leading players in the digital travel landscape share this commitment to accuracy.
Another concern is that if the gen AI platforms are to improve, they will need to rely even more on information from providers who work hard to present factual information that saves money and time.
The Gen AI platform’s growing appeal as an influential source could jeopardise this process by impacting the ecosystem’s ability to provide high-quality data.
It’s a similar situation to what we see in the media ecosystem. When gen AI platforms use traditional media stories to create news summaries instead of presenting the publication’s links, they tend to divert online traffic away from publishers’ pages. This affects the business model that supports experienced journalists and content producers in creating quality work.
The process can lead to a downward spiral, with less quality information to train generative AI models, ultimately impoverishing the entire ecosystem.
Digital travel platforms offering high-quality services have an advantage over media companies, as users easily identify our value. Still, a similar logic applies. If generative AI platforms start diverting significant online searches away from digital travel services, they could weaken the entire ecosystem, potentially reducing the availability of quality information.
A market with robust companies offering verified and factual data is crucial for maintaining travellers’ trust in online providers.
Part of the solution lies in established digital travel platforms harnessing the power of gen AI and combining it with their databases and human expertise to improve productivity, customer service, and personalisation at scale without compromising trust.
We may also see increased collaboration between big tech companies and established digital travel platforms to provide travellers with the best of both worlds, creating a positive cycle that benefits all players and, more importantly, consumers.
Generative AI technology has the potential to offer significant benefits and is here to stay. But to fully realise its potential, we must ensure the online travel ecosystem remains robust, reliable, and competitive.
Written by: Yesh Munnangi, CEO and Managing Director of Rome2Rio
BIO:
*Yesh (Yeswanth) Munnangi has been the CEO of Rome2Rio since 2020 and has a wealth of business and product experience in Australia and Europe. He was instrumental in helping Rome2Rio’s parent company, Omio, scale into one of Europe’s leading transport booking platforms and has held multiple positions at the company. He has also worked at Amazon and PWC.
