The Assistive Tool Paradox in radiology education: AI awareness, misconceptions, and job anxiety among Indian radiography students

The increasing integration of Artificial Intelligence (AI) into medical imaging is transforming the practice of radiology. However, the readiness and perceptions of future professionals, particularly in high-demand healthcare settings like India, remain insufficiently explored. This study aimed to evaluate the level of awareness, depth of technical understanding, and professional attitudes of radiography students in South India toward the clinical adoption of AI. A cross-sectional, multi-institutional survey was carried out involving 316 undergraduate and postgraduate radiography students from government and private institutions across South India. Data collection was performed using an 11-item questionnaire adapted from the validated instrument developed by Arruzza et al., with minor modifications to improve clarity and contextual relevance for Indian radiography students. Responses were recorded using a five-point Likert scale. Statistical analysis was performed using non-parametric methods. The Kruskal–Wallis H test was used to compare AI awareness and perception scores across academic years, the Mann–Whitney U test was used to compare responses by gender, Spearman’s rank correlation was used to evaluate associations between variables, and ordinal logistic regression was performed to identify independent predictors of AI acceptance. A majority of participants (63.0%) reported being generally aware of AI as an emerging trend (Mean score = 2.62 ± 0.98 on a 0–4 Likert scale). However, a key misconception was identified, with approximately 44% of students incorrectly categorizing conventional systems such as Automatic Exposure Control (AEC) and auto-positioning as AI technologies. A notable Assistive Tool Paradox was observed: although 66.2% of respondents acknowledged AI’s role in reducing workload, 43.4% expressed concerns regarding its potential impact on job opportunities. In the unadjusted analysis, academic level was associated with AI awareness (Kruskal–Wallis H = 15.36, p = 0.004); however, this association was no longer statistically significant after Bonferroni correction (adjusted p = 0.1203). Male students demonstrated significantly more positive expectations regarding patient acceptance of AI ( p = 0.004). Furthermore, a strong interest in formal AI education was evident, with 68.7% of participants supporting its inclusion in the curriculum. Although students demonstrate good general awareness, their understanding of AI fundamentals remains limited. The coexistence of optimism and concern highlights the need for structured, competency-based AI education introduced early in training, promoting a balanced and responsible integration of AI into clinical practice.