What will ChatGPT mean for the travel industry?

Every once in a while, a tech innovation develops seemingly overnight – and suddenly everyone, everywhere is talking about it. Examples that come to mind include Clubhouse and TikTok. But the pace at which ChatGPT has gone from unknown to ubiquitous seems unprecedented.

Launched on November 30 by OpenAI, ChatGPT (which stands for Generative Pre-trained Transformer) is a free artificial intelligence chatbot that has incredibly human-like communication skills (and a fee-based premium version in development). A Google search of the term “ChatGPT” turns up nearly 280 million results – with examples of its ability to write poetry, compose lyrics, play games, take tests and even code an entire website.

Actor and co-owner of Mint Mobile Ryan Reynolds demonstrates its use in a new ad for the mobile communications company. Microsoft is reportedly considering a $10 billion investment in ChatGPT owner Open AI according to multiple media outlets. And social media is saturated with posts arguing why ChatGPT is or isn’t the most powerful tool in the world.

So what – if anything – will ChatGPT and other generative AI chatbots mean for the travel industry? Will companies and individual operators use this technology in meaningful ways or will it fade into the background once the hype dies down (see past reference to Clubhouse). 

PhocusWire has reached out to a handful of brands and individuals for their perspectives. Some of the comments have been edited for brevity.

In addition, PhocusWire hosted its first LinkedIn Audio event on the topic of ChatGPT. Listen to the full 40-minute discussion with input from Expedia Group's Rajesh Naidu, Microsoft's Shane O'Flaherty and Kayak co-founder Paul English.

Rajesh Naidu, senior vice president and chief architect at Expedia Group

Recent breakthroughs with ChatGPT show that enormous opportunities still exist for technology to solve real world problems. For a long time, AI has been approached as a technology in search of a use case, and we’re starting to see this trend turn around as economic conditions put pressure on tech leaders to demonstrate a real return in investment for integrating AI. Because of this, tech companies are seeing value in data-driven personalization solutions and ChatGPT fills that need. It can help save companies time and money to allocate resources to solve more complex and specific problems.  

AI has ushered in a new era of personalized experiences, which ChatGPT can make even more sophisticated. AI tools need large data sets to feed off and our platform is massive; we have over 70 petabytes of data that powers six billion AI predictions annually. With a large dataset already in place, integrating this technology into our platform could hyper-personalize search results for travelers and supports our vision of offering open-ended, flexible search options. 

Conversational capabilities have come a long way over the past decade, but there is still a huge opportunity to improve on it by training chatbots to understand and predict traveler preferences. In just a few years, our platform has powered over 29 million virtual conversations, which saved more than eight million hours in agent time, allowing travelers to resolve issues faster with self-service. 

Another real opportunity for ChatGPT in travel is trip planning. Imagine if travelers could use it to create a trip itinerary or identify top hotels for their trip and have AI automatically add these recommendations to their Expedia trip board. It can simplify an extremely manual planning process involving sifting through lots of options down to minutes.

Arjan Dijk, senior vice president and chief marketing officer at Booking.com

ChatGPT really captured everyone's attention when it launched, but the technology is still in a relative state of infancy - it's already dividing opinion with responses that can range from entertaining and creative to glaringly inaccurate. There are some interesting potential use cases should its efficacy evolve and, as with all emerging and fast-growing technologies, it's certainly something we'll potentially look to explore, especially if we see that it's adding value to our customers' experience.

[Regarding marketing use cases] Marketing is at its best when it combines both science and art. We look to technology, data and the rigor and precision those can bring to guide us, but there is still much to be said for the creativity and intuition of our experts. It's too early to say whether the likes of ChatGPT can truly emulate those very human instincts. So for now, we're watching with interest, but it's not something I'd expect to see informing our work any time soon. 

We don’t experiment with technology for its own sake: it must actively solve a customer problem, remove friction and make it easier for everyone to experience the world - otherwise it’s not something we spend our time and energy on. That said, machine learning and artificial intelligence are already very much part of the Booking.com customer's journey, and these innovations will continue to pave the way to seamlessly connect all aspects of the travel experience. From AI-powered instant translation services that make it easier for travelers to negotiate last-minute changes of plan directly with accommodation hosts and rental car providers who don’t speak the same language to machine learning models that automatically inform our host about a late arrival to their beach cottage due to a delayed flight, technology will continue to smooth out the unforeseen bumps in the road with increasing finesse and proactivity. Whether ChatGPT or similar innovations add to that experience remains to be seen!

Patrick Surry, chief data scientist at Hopper

I think it shows we’re on the cusp of a significant change in the way society interacts with computers, which will lead to big changes in work, education and commerce. We think it has the potential to impact how we interact with our own customers. There’s some interesting use cases that we hope to explore such as improving automation and responses for common customer service requests through chat (and eventually even speech and video), as well as concierge-like services to recommend travel destinations, create personalized itineraries and the like.

It will also have a big impact on how we work, particularly in product development as we learn to use these new tools to improve the speed and quality of delivering our products and increasing our own efficiency and productivity. We will be able to experiment with how this can assist software development, as well as other use cases like making job descriptions more attractive to specific audiences. And I suspect this is just the beginning - we’ll need to go back to first principles and rethink a lot of how and why we do things the way we do. This will mean big opportunities for organizations that can adapt quickly and embrace change; something we’ve always prided ourselves on at Hopper.

Peter Syme, founder of 1000 Mile Journey

There will be a big impact on productivity and content production. You can’t deny that. Every single tour operator, even single hotel, every transport company now has the same opportunity to be as productive as a company much bigger than themselves. I see that as a commodity – everyone has access to it therefore, in some ways, where’s the value. But it’s the most productivity-enhancing thing I’ve ever seen. 

However I don’t think that’s the real impact. I think that’s just the start of the game. Once AI starts to get into the actual experiences – not just tour experiences but hotels, airports… the ability to speed up and enhance the actual experience, especially in something that is not a good experience at the moment like an airport, is going to be quite dramatic.

And the real game changer... is once the general public has it, it changes everything. You are a tour operator, and you create itineraries. The consumer can now do exactly the same thing as you and do it in real time. And 1,000 people can create 1,000 different experiences tailored to what they like. So the creation of a travel experience on the ground, which tour operators have dominated, is going to be disrupted by travelers being able to do it themselves. 

The weakness at the moment is it’s only scraping data through 2021. So it’s not the live data, it’s not up to date, the facts may not be 100% and it doesn’t have geolocation. It can give you an itinerary that doesn’t make sense. But take companies like Tripadvisor, Google - all the people with that data can put a wrapper around it, connect facts and geolocation and bang. I’d imagine Tripadvisor’s strategy team is awake all night at the moment trying to figure out if we’ve got all the unique data, how do we connect it with the power of this and then how do we monetize it. 

This could make entirely new businesses or it can destroy businesses. The speed this thing is developing is a thousand times faster than the internet developed. And all these people using it is speeding its development. And if the internet proves anything to us, it’s that once consumers find something really useful and that makes daily life better, they stick with it. So there are threats and opportunities there.

Matt Barker, founder and CEO of Horizon Guides

So far I've seen early adopters grabbing it with two hands, as well as the laggards who are barely aware it exists. For the early adopters there are certainly big opportunities but also some big risks. I'm working with a cruise brand who has decided to use ChatGPT for virtually all their new website content and will save hundreds of thousands of dollars on freelance costs, so there's big disruption to come. The written quality of the output itself is good and will get better, probably better than most people can write. For basic service copy – simple product descriptions, basic articles, email campaigns, that sort of thing, it's going to be very useful for brands and very painful for content creators.

But there are a couple of big caveats. First, people need to recognize that it's not generating anything new as much as scraping and regurgitating the internet. So there are issues with accuracy and truthfulness, and it needs human expertise to fact-check the output. This also means it can't add any novel details or insights, which is important for lots of content formats. The biggest risk is how Google responds, and if they'll accept AI content in the results. I doubt their ability to identify AI content in the long-run, but they can certainly further prioritize verified human expertise as a ranking signal. So you may not get “penalized” for using AI content, but you could be boosted for using expert human authors who can provide novel insights.

As far as Horizon Guides is concerned, you might think AI is an existential threat to a specialist publisher that invests heavily in new content creation. In actual fact we're doubling down on human contributors, and will soon be launching a Q&A feature for readers to ask our contributors questions directly. ChatGPT might herald a revolution in automated content creation, but I also think it could create a new premium for genuine human expertise too.

Alex Bainbridge, founder and CEO of Autoura

Fundamentally, my main thought is, I find the scope of what we have to think about quite intimidating, and that is coming from someone who has been working full time on the impact of AI on local tourism since 2018! Mainly this is because I don't really understand the core logic that is underneath all this generative AI. I graduated 25 years ago with a degree in computer science, all changes that have happened since have been understandable within the scope of what I learned then as an undergraduate. Now as a digital entrepreneur I am thinking that I don't really get this AI technology as much as I need to, in order to innovate with it. For the first time in my life I understand what it feels like to be made obsolete by a new technology.

My secondary main thought is, if you can generate infinite itineraries (for tours), you need to be able to operate infinite itineraries (for tours). That requires digital tour operating, which is where (long term) autonomous vehicles come in. (I call this "tour operating AI"). That will be the breakthrough that makes generative AI a consumer scale product in tourism & hospitality.

My third main thought is that personalization is now a high priority requirement, in order to make generative AI usable within tourism. This is where SSI comes in - we need to collectively share preferences over the industry, otherwise tourists are going to have a cold start with every AI they interact with. The hospitality group from DIF now has an important purpose.

Max Starkov, hospitality and online travel tech consultant and strategist

ChatGPT has been trained with information existing until end of 2021. Travel is a super dynamic category where this morning’s information is no longer relevant a few hours even minutes later. Inventory availability, prices, category of rooms/seats/cars change by the minute. You need good old-fashioned technology like CRS, WBE, RMS, etc. to handle these dynamics, not ChatGPT.

On the other hand you have chatbots in the industry that are highly trained and have been providing customer service, issue resolution and reservations on OTAs, Airline and major hotel websites for years. These chatbots are also AI- powered, highly specialized and self-learning, and much smarter and experienced than ChatGPT. So if I were a travel consumer, I will play for fun with ChatGPT, ask some silly questions, but if I want to dream about, research, plan and book travel, I will go to a travel brand or OTA site. 

A very heated discussion is raging on LinkedIn about whether the launch of ChatGPT marks the beginning of the end for Google’s near monopoly on finding the right answers about anything. Google is already the most-advanced AI company in the world, actually much better than any other company or government on this planet. Their search algorithm has been using AI for more than a decade now. In a way, Google already provides an AI-powered chatbot via its Google Answer Box, which provides quick answers to questions (not silly ones!) without the user needing to click to read further. 

Google can provide a much better AI-powered chatbot than ChatGPT that spews much better answers in milliseconds, but how are they going to make money? This is Google’s main problem today, not the AI technology. Charge users a fee every time they use the AI chatbot? Charge them a monthly subscription fee? There is a general “subscription fatigue” as many streaming and other services have discovered lately.

Dave Goulden, vice president of product at Sojern

AI can do amazing things and ChatGPT is a great example of that. For travel, there are some exciting use cases in terms of customer service, trip planning, and personalized customer experience. For example, it could help travel businesses analyze customer feedback and identify trends or patterns, providing valuable insight into customer sentiment and opportunities for personalization.

We think of ChatGPT as a fantastic example of where AI is heading to be more conversational, but there is a lot that we already do with AI. For example, Sojern leverages AI within our platform to drive marketing performance and personalization for users. In short, we process vast amounts of travel search and booking data, and then distill it into models of user behavior that we can use to predict future travel search and booking behavior. It’s what powers our clients’ success and continued growth. We’re able to provide travelers with relevant information based on what they’ve booked previously -- that level of scaled personalization isn’t possible without underlying AI to make sense of the data. 

Specific to ChatGPT, I believe that this technology will eventually be used in conjunction with a brand’s own content, in addition to the internet as a whole, so that it can be trained and customized by the brand to answer questions.

Mike Coletta, manager of research and innovation at Phocuswright

I’m very interested in seeing how travel companies leverage Generative AI for all the obvious uses in the coming year (marketing copy, sales copy, itinerary generation, customer service, optimization of backend processes) but like with most new breakthroughs, I’m most looking forward to the uses no one has even thought of yet. The longer-term implications of a technology like this are very difficult to anticipate

What’s also really interesting is how this once again demonstrates that the most disruptive innovations often come from outside the travel industry. As an example, for itinerary generation there is no shortage of AI travel planning tools, but it’s difficult if not almost impossible for them to gain trust among a critical mass of relatively infrequent travelers. ChatGPT is gaining that trust in its ubiquity and daily usage, so I think travelers could be much more likely to accept its itinerary recommendations without nearly as much research and shopping around. This has implications for human agents too.