How accurate are AI models in capturing Gen Z travel behavior?

A new study offers the first direct comparison between survey-based tourism data from Romanian Generation Z and artificial intelligence-generated insights, revealing sharp discrepancies in travel behavior interpretation. Titled “Exploring Tourism Experiences: The Vision of Generation Z Versus Artificial Intelligence”, the research was published in Administrative Sciences. It investigates how Romanian Gen Z travelers plan, experience, and reflect on travel, while evaluating whether AI models like ChatGPT can accurately replicate or predict their behavior.

The study, which utilized a mixed-method approach combining a 399-person online survey with AI-generated predictions, reveals that while AI can offer globally reflective insights, it often misses key cultural, economic, and regional factors that influence real-world behavior. The findings have serious implications for tourism professionals and marketers relying on AI to understand or target Gen Z consumers.

How accurate are AI models in capturing Gen Z travel behavior?

The research team first conducted a structured online survey targeting Romanian individuals aged 18 to 29. They explored travel frequency, transportation preferences, motivations, accommodation, dining, technology use, and entertainment. In parallel, they fed identical questions to ChatGPT-4, using an evaluation framework based on question-answer (QA) accuracy and source traceability. The AI’s responses were drawn from 12 global online sources, including industry blogs, travel reports, and research articles.

AI succeeded in matching 43.75% of the survey’s results. For instance, both AI and the survey accurately identified that Gen Z travelers frequently go on 2–3 trips per year and prefer planning their travels digitally. Both methods also confirmed that Romanian Gen Z does not typically use traditional travel agents, opting instead for online platforms or independent planning. They also found common ground in preferences for small group travel and the importance of natural destinations.

However, 56.25% of the AI-generated answers did not align with actual survey data. Notably, AI wrongly predicted air travel to be the most popular transportation method, while Romanian Gen Z respondents overwhelmingly preferred cars, likely due to cost, distance, and infrastructure. Similarly, AI assumed private accommodations like Airbnb were preferred, while the survey showed that commercial lodging with board included ranked highest. These mismatches underscore the AI model’s limitations in capturing localized behaviors shaped by specific economic or cultural conditions.

What do survey results reveal about Romanian Gen Z travel preferences?

The survey data depict a cohort that is digitally native but budget conscious, socially motivated, and deeply influenced by peer recommendations. Financial constraints were cited as the top barrier to travel by 77.78% of those who hadn’t traveled in the last 12 months. This economic awareness shaped other behaviors: most travelers avoided expensive transport options and valued price transparency when choosing services.

Romanian Gen Z favors small group travel with friends (39.85%) and shows significant interest in traveling with partners or family. Nature was a leading attraction, followed by visits to family or cultural sites. Dining preferences reflected a desire for variety and affordability, with multicuisine restaurants and fast food being favored more than immersive street food or high-end specialty dining.

On the digital front, Gen Z relies heavily on online sources to plan trips. Instagram (77.44%) and TikTok (61.65%) were the top platforms for destination research. However, the most trusted information sources were still personal - friends, family, and personal experience far outweighed influencers, blogs, or corporate social media channels. The most used tools included online booking platforms, official tourism websites, GPS maps, and review aggregators.

Their favorite leisure activities included independent exploration (82.71%), hiking (51.88%), and local shopping (51.13%). A preference for experiences over formal tours or scheduled activities was apparent. For accommodation, while commercial hotels led, a sizable portion still opted for private or familial housing.

Can AI accurately predict and replace traditional travel research?

The study concludes that AI has strong potential to support, but not replace, traditional tourism research. ChatGPT was able to provide generalized insights reflecting global Gen Z trends, such as eco-conscious travel, tech-reliant trip planning, and a preference for experiential tourism. However, it consistently failed to account for regional specificity and socioeconomic nuance.

AI's overreliance on global content, often skewed toward Western travel norms, leads to predictions that may be misaligned with realities in markets like Romania. The authors emphasize that tourism professionals must use AI carefully, integrating it with localized data and direct consumer engagement. Blind reliance on AI risks strategic missteps, especially in market segments with high cultural variability or developing infrastructure.

Instead, AI is best used to complement human research, offering a broader analytical layer. By combining empirical local surveys with AI-derived global benchmarks, tourism organizations can achieve a balanced view that is both scalable and context-sensitive. Smart destinations, the study suggests, should train AI models with localized datasets to improve contextual accuracy. This could help refine AI tools to better reflect the nuances of regional consumer behavior, rather than reproducing global averages.

The authors also highlight ethical concerns: AI tools are prone to bias based on their source material, often amplifying content from luxury travel platforms or influencers. This can distort the real preferences of economically constrained or culturally distinct demographics. Transparency, customization, and critical evaluation must be built into any AI-driven decision-making system.