The AI revolution is coming to travel
Here & There - 02.15.23
You’ve probably heard of ChatGPT and GPT-3 at this point. I’ll be honest — this new AI technology is incredibly impressive and advancing at an insane rate. People are using these technologies to write code, do interview prep, re-write resumes, tell jokes, create custom bedtime stories for their kids, create customized meal + workout plans, and a million other things.
Here’s what ChatGPT has to say about itself:
“ChatGPT is an artificial intelligence language model that has been trained on a large amount of text data to understand and generate human-like responses to text-based questions or prompts. In simpler terms, it's like a computer program that you can have a conversation with, and it will understand what you're saying and respond to you as if it were a person.
ChatGPT uses a technique called "deep learning" to analyze patterns and relationships within large datasets of text. By doing so, it has learned to recognize context, understand meaning, and generate coherent and natural-sounding responses to a wide variety of questions and prompts.”
The number of projects launched on Product Hunt in the last ~30 days that use or are somehow tied to ChatGPT is truly astronomical. ChatGPT and GPT-3 (and other AI-based solutions) will inevitably become part of the travel and outdoor experience.
Where it won’t be used (yet)
ChatGPT still needs to be trained on an existing dataset, so more timely, up-to-date news and journalism is still in the purview of humans for now. It can’t write your personal experiences, recap an expedition that just happened, or dive into the details of the latest gear release. Outdoor/travel journalism isn’t going away altogether, but it’s definitely going to change.
Where it will be used
SEO/Listicles/generic travel content
This type of stuff is square in the sights of AI. You know the kind of articles I’m talking about. They probably start with “5 best” or “10 places” and basically rehash information and content that has been created on a million different websites and blogs already. Nobody put too much effort into them to begin with, so it’s not surprising that AI will write extremely passable articles. These types of articles an extremely powerful SEO tool, since they line up with high volume searches like “best hikes near [location]”. While companies will definitely want to use AI to create vast amounts of content with less work, what we don’t know yet is how Google plans to treat this type of content. Will they sink “AI generated” content in the search results (if they can detect it)?
Curation. Curation. Curation.
There is *so much* outdoor/travel content out there that many people prefer to search Google for lists like I mentioned above rather than slog through a million options in a search experience on Alltrails/TripAdvisor/etc. Some companies, like Fora and Thatch are already tapping into this need for travel curation (although not with AI, yet).
I think we’re likely to see AI being used as a curator/travel agent of sorts — building itineraries, generating suggestions, and adapting based on user preferences. It might either be the primary experience, or used by human curators to dramatically speed up their work. Here’s an interesting demo shared by Tripnotes (pre-launch):
Customization and learning.
Something ChatGPT does quite well is adapting its answers based on parameters, or “acting” as if it has a particular persona. In this way, you could do some interesting things with personalization of travel suggestions, or perhaps anthropomorphize several personas of ChatGPT to give you suggestions based on how you plan on traveling. Here’s an example.
I asked ChatGPT to suggest a few hikes for me near Aspen, Colorado. It gave me:
Maroon Lake Scenic Trail (1.9-mile loop)
Cathedral Lake Trail (5.4 mile out-and-back)
Hunter Creek Trail (8.9 mile out-and-back)
Smuggler Mountain Trail (14.0-mile out-and-back)
These are all short-medium length, popular trails near Aspen and decent suggestions for the average visitor. However when I asked ChatGPT to treat all following questions as if it was having a conversation with a “fit, early 30’s ultra runner who is comfortable with long days on the trail and challenging elevation gain”, the response changed to:
Maroon-Snowmass Trail (21.3-mile out-and-back)
Capitol Creek Circuit Trail (24.6 mile loop)
Four Pass Loop (28 mile loop)
Conundrum Hot Springs Trail (17-mile out-and-back)
Now, you might be saying “well, it basically just changed the distance parameters”. And for this example, you might be right. But more importantly to me, these are the trails I would be interested in. But, you could build on the above persona with your personality, and it could adapt suggestions for different types of hikes, local cultural attractions, dog-friendly hikes, or your favorite types of nearby food, or other activities, etc. Honestly, using it for outdoors-focused travel is probably a narrow view – places that have exponentially more overwhelming choices for activities/food/attractions (like cities) will be where something like this really shines.
Prepare for every travel site to use an AI chatbot.
Real time translation. Training instances of ChatGPT off of information from *real* local experts. Chatbots for destination websites. Personalized packing lists based on your itinerary, local weather, etc.
Because GPT-3 is available to anyone who pays for API access, this means that we’re likely to se a lot of similar experiences both newly created, and integrated features into existing platforms. How might platforms or apps that all use GPT3 for travel suggestions differentiate themselves?
Design + ease of use. Design is always something that can make an app stand apart. An experience utilizing GPT3 still needs to have an intuitive and understandable UX and provide an experience that users will enjoy coming back to.
Uniqueness of data. When the underlying technology becomes largely the same, the uniqueness of your data and content becomes a key factor. Who has the best images? The most in-depth reviews? Who has great proprietary information vs who is just acting as a presentation layer for data scraped from google? Which platforms are providing *great* suggestions or supplementary data due to some amount of human curation, vs the generic *most popular* suggestions?
Trust + Brand + Community. These aspects arguably become more important than ever. When anyone can spin up a ChatGPT powered company/feature/app basically overnight, customer alignment, community investment, and trust in your brand becomes vital. Do you believe in the mission of this company (regardless of whether it’s AI driven or not)? Do you feel more connected to a company that is still “human driven”. Do you feel connected with other people using this product?
But what would ChatGPT say?
After I finished this article, I used my above question “…how might platforms that use ChatGPT differentiate themselves?” and posed it to ChatGPT. Here’s what it said 😲.
That's a great question! As ChatGPT becomes more widely available and integrated into travel platforms and apps, it's likely that we'll see a proliferation of similar experiences and features that use the technology to make personalized travel recommendations. To differentiate themselves in this crowded market, platforms and apps may need to focus on a few key areas:
User experience: Platforms and apps that use ChatGPT can differentiate themselves by creating a user experience that is intuitive, engaging, and enjoyable. This might include features such as gamification, social sharing, or personalized content that keeps users coming back.
Data sources: Platforms and apps can differentiate themselves by using unique data sources to make travel recommendations. For example, a platform that integrates data from social media, weather apps, or local news sources may be able to provide more accurate and timely recommendations than a platform that relies solely on standard travel data.
Niche focus: Platforms and apps can differentiate themselves by focusing on a specific travel niche or interest, such as adventure travel, sustainable tourism, or cultural experiences. By tailoring their recommendations to a specific audience, these platforms and apps can provide more relevant and personalized suggestions that stand out from more general travel apps.
Additional features: Platforms and apps can differentiate themselves by offering additional features beyond travel recommendations, such as booking tools, itinerary management, or virtual reality experiences. By creating a more comprehensive travel ecosystem, these platforms and apps can provide a one-stop shop for all of a traveler's needs.