Survey: Reliability Mattered Most To Travelers Letting AI Plan Trips

Reliability Beats Transparency in Travelers’ Willingness to Let AI Plan

Letting an AI app plan a vacation is supposed to save time and mental energy. Yet in a new survey of 421 travelers, people who wanted more openness from these tools about their data and design were less willing to hand over their travel decisions, a surprising result, since AI research has repeatedly stressed the need for transparency.

Researchers had predicted the opposite. Their starting hypothesis was that travelers would be more willing to hand decisions to AI tools they saw as transparent. Instead, reliability, meaning accurate and consistent recommendations, and accountability, meaning someone answerable when a trip goes wrong, were the qualities consistently tied to willingness to let AI take over planning.

Published in Tourism Management Perspectives, the study looks at a field where travel companies are adding AI planners. Tripadvisor has reportedly launched an automated itinerary service, Expedia uses AI language processing to support more than 30 million virtual agent conversations, and the paper cites a forecast that the global market for generative AI in travel is expected to grow from about $894 million in 2024 to roughly $5 billion by 2034. Earlier research mostly focused on how useful AI tools are, while this study asks when travelers are willing to hand over the decision itself, which has become a business question.

Reliability and Accountability Held Up at Every Level of AI Travel Planning

A team from Edith Cowan University in Australia and two other universities surveyed 421 travelers recruited through an Australian online panel company. All had previously handed travel decisions to generative AI, the type of chatbot that writes answers on demand, and the survey included attention-check questions to verify response quality. The paper describes the group only as having diverse backgrounds, with no age or gender breakdown. Australia was chosen because it ranks among the world’s frontrunners in generative AI use.

Participants rated how strongly they agreed with statements about AI on a seven-point scale. Five qualities were scored: reliability, fairness, trustworthiness (belief that an AI acts with integrity in the user’s interest), accountability, and transparency. The transparency statements asked whether AI makers should be clear about how old a system’s data is, where it came from, how the system was tested, and who designed it.

Travelers then rated their willingness to delegate at three levels. At the lightest, AI decides what should matter when planning a trip. In the middle, AI trims a long list of options to a few, and the traveler picks. At the deepest, AI makes the final call, with travelers rating agreement with statements such as “GenAI should make the final decision when planning my travel, not me.”

Reliability grew stronger as the stakes rose, peaking when AI made the final decision. Accountability held steady across all three. Trustworthiness mattered on its own only at the lightest level, when AI was merely helping decide what to weigh. Fairness, whether an AI seemed unbiased, showed no measurable link at any level, which the authors suggest may stem from bias in an algorithm being mostly invisible to users. Together, the five qualities accounted for roughly a third of the differences in willingness, leaving much of what drives travelers’ choices unmeasured.

Higher Transparency Ratings Went With Less Willingness to Hand Over Trips

Transparency ran the opposite direction at all three levels. Travelers who agreed more strongly that AI should be open about its data and design were less willing to delegate, though the link was modest, smaller than reliability’s.

As a possible reason, the authors point to “cognitive overload,” meaning more information than a person wants to process. People hand off planning to save effort, and weighing an AI’s reasoning hands some of that effort back. Because the survey captured opinions at one point in time rather than testing different AI explanations on travelers, it shows a link, not a cause. Another untested possibility is that travelers who care most about how an AI works are simply more cautious, which alone could explain their hesitation.

A second analysis, which examined combinations of qualities instead of one at a time, added detail. The only combination that predicted handing over final decisions was high reliability, fairness, and trustworthiness with low transparency, even though fairness showed no effect on its own. High transparency also appeared in most of the combinations linked to low willingness at all three levels, yet it showed up in some that predicted willingness at the lighter two levels, so its downside was not universal.

Researchers Say AI Travel Planning Explanations Should Be Optional

Based on the results, the authors advise developers to invest in accuracy and clear accountability, such as feeding the AI up-to-date travel information, rather than focusing only on explanations. They recommend that transparency be “a user-controlled factor rather than a system-imposed information overload.” Those recommendations are the authors’ own suggestions, not approaches the study tested.

Travel companies counting on more transparency to win customers over may be solving the wrong problem. In this survey, perceived reliability had the strongest tie to willingness to hand over the trip.

---

Disclaimer: This article summarizes a single survey study for general informational purposes only. Its findings reflect traveler opinions in one country and show associations, not cause and effect. Nothing here is travel, business, or professional advice, and readers should consult the original paper for full details.

---