If you have difficulty in submitting comments on draft standards you can use a commenting template and email it to admin.start@bsigroup.com. The commenting template can be found here.
Standards Development
This document provides guidelines for the application of Artificial Intelligence (AI) techniques to semantic data for tourism destinations structured in accordance with ISO 20525 (Tourism and related services – Semantics applied to tourism destinations).
It covers two major application areas, (i) the consumption of this type of semantic data by AI systems, including recommender systems, conversational agents, search and retrieval engines, and decisionsupport tools, and (ii) the generation or enrichment of this type of semantic data by AI-based processes, including natural language processing (NLP), machine learning (ML), and generative AI. These processes can be used to create, annotate, validate or enhance tourism content in conformance with the semantic structures defined in ISO 20525.
This document includes guidelines and recommendations for ensuring the quality, consistency and traceability of AI-processed semantic tourism destinations data, specifically:
Semantic data consumption guidelines to ensure the correctness of the information supplied by AI systems.
Interoperability recommendations for AI systems that exchange semantic tourism destinations data across platforms and organizations.
Human oversight and governance considerations for AI-generated semantic tourism destinations data.
Guidelines to preserve the integrity of the tourism destinations ontological model when consuming or generating data.
Considerations to ensure the transparency and explainability of AI decisions based on semantic tourism destinations data.
The following is outside the scope of this document:
The semantic data structures and ontological model for tourism destinations defined in ISO 20525. Requirements for establishing, implementing, maintaining, and continually improving Artificial Intelligence Management Systems (AIMS), addressed in ISO/IEC 42001. Guidance on risk management related to AI, addressed in ISO/IEC 23894. Personal data protection and privacy in tourism systems. Tourism AI applications that do not interact with semantic data structured in accordance with ISO 20525.
1. Market need
The adoption of AI across the tourism sector creates both opportunities and tangible risks. The tourism industry is deploying all sorts of AI systems to consume, process, and generate tourismrelated structured content. However, without standardized guidelines grounded on solid semantic structures, the quality of AI-generated or AI-processed data may suffer from a set of important problems related to AI systems, especially hallucinations (data inconsistent with reality), a lack of traceability to link statements produced by an AI-system and its sources, and a lack of data integrity (e. g. incomplete information supplied to the tourist).
Major tourism platforms (including online travel agencies, tourism destination management systems, and booking engines) are integrating AI-powered recommendation and content generation capabilities as a key strategic and competitive priority. This trend is accelerating globally, in both public and private tourism sectors.
The nature of the problems in a sensitive business context like tourism, where trust is paramount, can ruin the credibility of tourism destinations if their AI systems are not managed properly. Thereby the need to address these issues.
2. Problem this standard solves
A powerful way to ensure the reliability of tourism destinations AI systems is to make them interact with semantic tourism data, which strongly encode tourism content (such as point of interest description, media, and service properties) in a model defined within the ISO 20525 standard.
This standard addresses the following risks:
Semantic hallucination risk: LLMs and generative AI tools are known to produce factually incorrect or internally inconsistent structured data, a risk that can materialize while enriching tourism databases and knowledge graphs.
Misinterpretation of ontological relationships: AI recommender systems consuming semantic tourism destinations data may misinterpret hierarchical or associative ontological relationships if no consumption guidelines are defined, leading to poor-quality, inaccurate, or misleading recommendations for tourists.
Lack of transparency and human oversight: AI recommender systems can produce biased judgments not grounded on factual data, eroding tourist trust and exposing destinations to reputational and regulatory liability (e.g. recommending a dangerous hike trail to a group of unprepared people).
3. Gap analysis
ISO/IEC JTC 1/SC 42 develops general-purpose AI standards (ISO/IEC 42001, ISO/IEC 23894, ISO/IEC 22989) applicable across all industries.
ISO/TC 228/WG 21 develops tourism-specific semantic standards (ISO 20525). Neither committee addresses the specific interface between AI systems and tourism destinations semantic data. National precedents exist (UNE 178503 in Spain), but no international standard bridges AI systems and the tourism ontological model at the operational level that this proposal targets.
Required form fields are indicated by an asterisk (*) character.
You are now following this standard. Weekly digest emails will be sent to update you on the following activities:
You can manage your follow preferences from your Account. Please check your mailbox junk folder if you don't receive the weekly email.
You have successfully unsubscribed from weekly updates for this standard.
Comment by: