AI Personalization in Tourism Needs Trust, Not Just Efficiency

Artificial intelligence is rapidly reshaping how tourists choose destinations, plan itineraries, and experience places. AI stands as a ‘beacon of transformation’ in travel,” observes the World Economic Forum, highlighting how AI can dramatically reshape tourism planning and experiences. From chatbots that suggest restaurants to recommender systems that tailor cultural visits, AI-driven personalization is becoming a standard feature of digital tourism services. Policymakers often frame these tools as neutral efficiency gains. The assumption is simple. If AI is useful, people will adopt it. Indeed, traditional models like the Technology Acceptance Model (TAM) emphasize that perceived usefulness is the primary determinant of technology acceptance. 

Evidence from the field suggests otherwise

Based on empirical research conducted in Marrakech, Morocco, this piece argues that trust, not technical performance alone, is the decisive factor shaping whether AI personalization creates long-term value for users. This insight matters for current debates on AI governance, especially in sectors where AI directly mediates human experience and decision-making.

Why is usefulness not enough?

Technology policy has long relied on models that prioritize usefulness and efficiency. This logic is embedded in many regulatory discussions around AI adoption. If systems perform well, adoption will follow. 

Yet in tourism, AI does not simply optimize tasks. It shapes emotions, expectations, and perceptions of a place. 

In a survey of 282 tourists using AI-based tools such as chatbots, virtual guides, and personalized recommendations, ease of use mattered more than perceived usefulness. Tourists were more satisfied when AI felt intuitive and adaptive, not when it merely delivered correct information. Perceived usefulness did not directly increase satisfaction.

This finding challenges a common policy assumption. In experiential sectors, AI success depends less on functional output and more on how users feel during interaction. AI-driven experiences, especially storytelling or personalized recommendations, can shape how tourists feel, imagine, and interpret a destination rather than just what they know about it. Systems that optimize efficiency but ignore user perception risk underperforming in real-world settings.

This has implications for how regulators define high-risk and general-purpose AI systems. User-facing AI should not be evaluated solely on accuracy or productivity metrics. Consumers are enthusiastic about personalization but wary of privacy: a 2025 IAB survey (reported at a Tourism Innovation Summit) found that 82% of travelers appreciate personalized ads, yet 80% worry about data misuse. This highlights the “privacy–personalization paradox” in tourism technology. Experience-based sectors require additional safeguards and design standards that account for trust and perceived agency.

Trust as a policy variable

Trust emerged as the strongest predictor of loyalty in this study, even more influential than satisfaction. Tourists who trusted AI systems were more likely to reuse them and recommend them, regardless of performance measures. 

Importantly, trust did not function as a simple modifier between satisfaction and loyalty. Instead, it operated as a foundational condition. Users who did not trust the system disengaged early, before satisfaction could even form.

This distinction matters for AI governance. Many regulatory frameworks treat trust as an outcome of compliance. Indeed, many AI governance efforts focus on technical standards for performance (e.g. accuracy, bias) and categorize applications by risk level, but they often treat user trust as an afterthought. For example, UNESCO’s AI ethics recommendation highlights transparency, explainability, and human oversight as core principles. The OECD AI Principles call for “innovative and trustworthy AI” that includes human oversight. The EU’s new AI Act similarly states its aim to “ensure that Europeans can trust what AI has to offer” by enforcing data protection and bias mitigation.

In practice, trust functions as an entry condition. Transparency, data governance, and explainability are not optional ethical add-ons. They determine whether AI systems are used at all.

In tourism, this includes clarity about data use, limits of recommendations, and cultural sensitivity. AI tools that personalize experiences without explaining how or why risk being perceived as intrusive rather than helpful. UNWTO stresses ethical AI in tourism, noting the need to balance AI with authentic human interactions and data privacy. Destination organizations are advised to implement AI training and transparent data policies to maintain trust. 

Lessons from the Global South

Most AI policy debates are informed by data from North America, Europe, or East Asia. However, as one policy brief warns, without inclusive governance, the digital divide can become a “governance divide,” excluding billions from shaping AI’s future. In fact, less than 10% of global AI governance frameworks originate in the Global South. Emerging markets are often treated as future adopters rather than present test cases. The Marrakech case shows why this is a mistake.

Brookings and others point out that AI policies must consider developing-world realities. In contexts where digital adoption is uneven and institutional trust is fragile, users scrutinize AI systems more closely. Trust becomes a central determinant of legitimacy. This makes Global South contexts valuable policy laboratories, not peripheral cases. For instance, in Morocco, only ~60% of people (and just 40% in rural areas) have digital access,  and reliance on foreign algorithms can raise bias concerns. Governance thus needs local adaptation. 

AI governance frameworks that fail to incorporate these realities risk exporting systems that function technically but fail socially. For international policy discussions, this means shifting from a one-size-fits-all approach toward context-aware regulation.

Policy implications

If AI personalization is to be responsibly deployed, especially in consumer-facing sectors, policymakers should integrate trust-based criteria into AI oversight. This includes mandating transparency standards, supporting human-centered design, and recognizing experiential impact as a governance concern.

AI systems do not operate in a vacuum. They mediate relationships between users, services, and institutions. When trust is absent, efficiency does not compensate. Good AI policy must therefore regulate not only what systems do, but how they are experienced.

Author Bio:

Bendjedid Rachad Sanoussi is an ICT specialist and PhD researcher in digital and artificial intelligence for management at the Euromed University of Fes, in Morocco. His research focuses on AI-driven personalization, trust, and user experience in sustainable tourism. He is currently a visiting researcher at CENTOURIS, University of Passau in Germany, where he works on digitalization and sustainable tourism transitions. He has contributed to AI policy analysis at the Centre for AI and Digital Policy (CAIDP) and previously worked with UNDP and ITU on digital transformation initiatives.