# Psychometric validation of the Turkish Brief General AI Self-Efficacy Scale: examining its role in AI anxiety, acceptance, and demographic predictors

## TL;DR

This study validates a Turkish AI self-efficacy scale and finds that higher self-efficacy reduces AI anxiety and increases acceptance of AI in education.

## Contribution

The study integrates self-efficacy theory with TAM to show how AI self-efficacy mediates anxiety and acceptance in pre-service teachers.

## Key findings

- Higher AI knowledge predicts greater self-efficacy and acceptance, and lower anxiety.
- AI self-efficacy partially mediates the relationship between AI anxiety and generative AI acceptance.
- The Turkish GSE-6AIS scale is reliable, valid, and gender-invariant.

## Abstract

The rapid integration of artificial intelligence (AI) technologies in education highlights the urgency of understanding pre-service teachers’ readiness to adopt these tools effectively. Although prior research has separately examined AI self-efficacy, AI anxiety, and generative AI acceptance, few studies have investigated their interrelations within a unified framework. This study linguistically adapted and psychometrically validated the Turkish version of the Brief General AI Self-Efficacy Scale (GSE-6AIS) and explored its associations with AI anxiety, generative AI acceptance, and demographic characteristics.

Data were collected from 941 pre-service teachers (52.4% female, 47.6% male; M = 21.97 years, SD = 1.45) recruited via convenience sampling from seven Turkish universities. Confirmatory factor analyses supported the scale’s unidimensional structure and internal consistency, and multi-group analyses indicated gender invariance.

The results showed that higher AI knowledge predicted greater self-efficacy and generative AI acceptance, and lower AI anxiety, whereas gender and computer use showed no significant effects. Mediation analyses revealed that AI self-efficacy partially mediated the relationship between AI anxiety and generative AI acceptance, highlighting its role as a key psychological mechanism. Integrating Bandura’s self-efficacy theory with the Technology Acceptance Model (TAM), findings highlight AI self-efficacy as a central mechanism linking anxiety to generative AI acceptance.

These findings indicate that the Turkish AI Self-Efficacy Scale is a reliable and valid measure and underscore the importance of fostering self-efficacy to reduce anxiety and enhance acceptance of generative AI in educational contexts. The results have practical implications for teacher education programs aiming to prepare future educators for the increasing presence of AI in learning environments.