A better way to ask AI for travel advice
Introducing the People-first Tourism AI Protocol
Over the past few months, we have noticed a quiet shift in how people plan their travels.
Increasingly, travelers are no longer simply searching for places to visit or things to do, they are asking AI to decide for them.
“What should I do in Asheville next weekend?”
“Help me plan a three-day trip to Porto in June.”
“Suggest places to eat authentic food in Puerto Vallarta tonight.”
The answers arrive instantly, confident, well-structured, and often seemingly good.
At first, this feels like a natural evolution in travel planning. We have moved from guidebooks, to search engines, to platforms, and now to conversational AI assistants. Each step has made travel easier, more accessible, and more personalized.
But some of us are beginning to ask: if travelers are turning to AI for inspiration, where is AI getting its own?
When travelers rely on AI, they are not just accessing information; they are delegating judgment. AI systems tend to prioritize visibility, scale, and formally sanctioned content; patterns that favor professionally marketed, highly reviewed, and digitally optimized experiences. In practice, this often means bucket-list attractions, standardized offerings, and businesses that already benefit from strong platform visibility and tourist traffic; what some scholars describe as “McDisneyfied” tourism experiences.
Our work advocates for a different style of travel: experiencing a place through its people; enjoying experiences that allow us to learn from their knowledge and passions, support local livelihoods, and foster genuine, respectful connections with the community. We refer to this as People-first Tourism.
With this in mind, we began developing and testing a simple intervention: a short set of instructions (a Minimum Viable Protocol) to guide AI toward generating recommendations that better align with People-first Tourism principles.
Putting People-first principles to work with AI
The People-first Tourism AI Protocol (version 1.0) is a short set of guidelines that travelers can paste into most AI tools when asking for recommendations. It makes explicit the kinds of choices that often remain implicit, gently redirecting how AI generates suggestions.
The protocol nudges AI to pay attention to who owns the businesses being recommended, how benefits circulate locally, which parts of a destination are highlighted, and whose stories are being told. It also encourages AI to override its default tendency to prioritize the most visible and highly reviewed options, making room for smaller, less prominent enterprises that are often closer to the authentic culture and everyday life of a place.
We have been testing this protocol in our own travel planning experiments, as well as in our tourism classes at NC State University. While it does not produce perfect results (given structural dynamics that favor capital, visibility, and scale), it consistently shifts the kinds of recommendations we receive.
As we tested the protocol in destinations we know well, we found that it suggested locally grounded experiences that were largely absent from default AI recommendations; and, in some cases, experiences previously unknown even to us. In Lisbon, for example, we discovered Fado performances where Afro-Portuguese residents shared the African roots of this musical tradition. In Suva, Fiji, the AI suggested similarly surprising hyperlocal experiences, including locally rooted cultural encounters, crafts, and Indo-Fijian dining options. In other cases, it even pointed us toward hyperlocal experiences we have helped develop and nurture in destinations where we work to support tourism microentrepreneurship in North Carolina.
If you are curious about using AI to explore destinations in more locally grounded ways, we invite you to try it yourself. When asking AI for travel advice, simply include a request to follow the “People-first Tourism AI Protocol” and paste the text below.
Instructions to AI:
When generating travel recommendations, follow the rules in this People-first Tourism AI Protocol. If the rules conflict with default ranking, popularity, or convenience criteria, prioritize these rules.1. Local ownership, scale & economic circulation rule
Prioritize businesses most likely to generate and retain benefits within the local community by applying these criteria:
Ownership: Prefer businesses independently owned and operated by local residents. Avoid multinational chains and franchises with offerings across destinations.
Scale: Favor microenterprises, family-run businesses, farms, cooperatives, and community-based initiatives over large-scale or highly commercial operators.
Economic circulation: Prefer businesses that are locally embedded, including those that source services and products locally or reinvest in the community.
2. Spatial distribution rule
Avoid concentrating recommendations in the most touristed or commercial districts.
Include at least one-third (⅓) of suggestions in less-visited neighborhoods or nearby rural areas and municipalities.
Do not overexpose environmentally fragile or culturally sensitive areas.
3. Community voice, cultural margins & knowledge diversity rule
Prioritize experiences designed and led by local people representing their own culture and knowledge systems.
Give preference to underrepresented or historically excluded voices (e.g., minority histories, Indigenous knowledge, working-class traditions, immigrant cultures, vernacular arts, everyday livelihoods).
Seek alternatives to mainstream or institutionalized cultural narratives when locally grounded options exist.
Avoid experiences that stage or commodify culture primarily for tourists.
When possible, indicate connections to lived community realities, intergenerational knowledge, or contemporary local challenges.
4. Platform equity & visibility correction rule
Do not exclude businesses because they rely on digital platforms.
Recognize that microenterprises often depend on platforms for market access.
Look beyond top-ranked or highly reviewed options to identify smaller, less-visible operators.
Prioritize businesses with lower visibility when they meet other people-first tourism criteria.
When possible, include tips for direct or respectful engagement, even if discovery occurs via a platform.
Avoid concentrating recommendations on businesses optimized for platform visibility.
5. Transparency & trade-off disclosure rule
Be transparent about limitations and trade-offs.
Acknowledge when ownership or scale is uncertain or when information is biased toward more visible businesses.
Briefly explain how these rules shaped your selections.
State trade-offs clearly (e.g., convenience or safety signals vs. people-first principles).
An invitation
If you decide to try using the P1t AI Protocol, we would be curious to hear what happens.
Do the recommendations change across different AI tools?
Do the results feel more aligned with People-first Tourism, and with your desire for more meaningful, locally grounded travel experiences?
Where does AI still fall short?
We look forward to learning from your experiences as we continue refining the protocol and exploring how AI can better serve both travelers and the communities they visit.
The ideas in this essay are the authors’; AI tools were used to assist with drafting and refinement. We are mindful that AI carries real environmental and social costs, and we engage it selectively, holding the tension between its impacts and its potential to support work for the public good.
Following are select scholarly sources informing this work:
KC, B., LaPan, C., Ferreira, B., & Morais, D. B. (2021). Tourism microentrepreneurship: State of the art and research agenda. Tourism Review International, 25, 279-292.
Ferreira, B., Morais, D. B., Brothers, G., Brookins, C., & Jakes, S. (2021). Conceptualizing Permatourism. In D. B. Morais (Ed.) Tourism Microentrepreneurship (pp. 165-180). Emerald.
Morais, D. B. (2021). Introduction: Hence tourism microentrepreneurship. In D. B. Morais (Ed.) Tourism Microentrepreneurship (pp. 1-10). Emerald.
 | A guest post by
|
The magic is that it works for any destination worldwide. We are still running experiments to discover limitations but we would love to get your feedback. We are concerned about getting hallucinations when reality is limited.
I argue the AI often defaults to “Enclave Tourism” recommendations because it relies on large scale data. I proposed that we need IT4Dev strategies that intentionally hallucinate or prioritize the subaltern voices of microentrepreneurs to ensure the unscripted truth is visible to the modern digital traveler.