System Hallucination Frequency and Service Failure Recovery in AI-Assisted Heritage Education: Psychological Safety, Collaborative Learning, and Technological Contingencies
Generative Artificial Intelligence (GenAI) is finding growing use in heritage education, but system hallucinations can impair learning and service recovery in culturally sensitive settings. The associations of System Hallucination Frequency (SHF) with On-Site Service Failure Recovery Effectiveness (SFRE) are examined through Trainee Psychological Safety (TPS) and Trainee Collaborative Learning Behavior (TCLB), with technological contingencies. Data were collected from 427 university students in Saudi Arabia with experience in AI-supported heritage-related learning using a cross-sectional survey. Structural Equation Modeling indicated that higher perceived SHF was negatively associated with SFRE, TPS, and TCLB, whereas TPS and TCLB were positively associated with recovery effectiveness. TPS and TCLB also showed significant individual and serial indirect associations. Moreover, TCLB and perceived retrieval quality (CRA-TKP) buffered the negative SHF-TPS association, and perceived flagging quality (AGFP-CCN) strengthened the positive TCLB-SFRE association. The results identify associations consistent with the proposed AI-mediated service-recovery model and offer practical guidance for designing reliable and culturally sensitive AI-supported heritage education systems.
Authors
- E. Álvarez (ORCID: https://orcid.org/0000-0002-0473-6037)
- Ahmed E. Abu Elnasr (ORCID: https://orcid.org/0009-0004-7338-8787)
- Ahmed K. Elnagar (ORCID: https://orcid.org/0000-0003-1787-8189)
- Abu Elnasr E. Sobaih (ORCID: https://orcid.org/0000-0002-2730-689X)
- Abdallah M. Elshaer (ORCID: https://orcid.org/0000-0002-7801-2111)
- Tamer M. Abdelghani
Institutions
- Suez Canal University (EG)
- Universidad de Granada (ES)
- University of Sadat City (EG)
- University of Khorfakkan (AE)
- King Faisal University (SA)
Publication Details
- Journal
- Algorithms
- Published
- 2026-10-07
- DOI
- https://doi.org/10.3390/a19100852
- Primary Topic
- AI in Service Interactions
- Type
- article
- Field-Weighted Citation Impact
- 0.00