ZAGREB, Nov. 21 (Xinhua) -- Most Croatian citizens remain cautious about relying on artificial intelligence (AI) for healthcare, education and major decisions, according to an HRejting poll released by Croatian Radiotelevision on Saturday.
Conducted from Nov. 13 to 17 among 1,000 respondents, the survey found that 41 percent of people have only superficial knowledge of AI, while 37 percent use AI tools such as ChatGPT. Ten percent use AI regularly at an advanced level, and 12 percent are unfamiliar with it.
Regarding the healthcare sector, over 60 percent rejected an AI-generated medical diagnosis, citing the need for human experience and empathy. About one-third accepted it only under a doctor's supervision, while roughly 3 percent would fully trust AI even if it proved statistically more accurate.
Similarly, 65 percent opposed AI replacing teachers, stressing the need for human contact in education. Roughly one-third viewed AI as a useful teaching aid, and only 3 percent believed it could fully replace teachers.
In critical areas such as traffic, finance, and the judiciary, 65 percent trusted humans over algorithms due to morality, intuition, and contextual understanding. Fifteen percent trusted both equally, about 13 percent trusted neither, and 6 percent favored algorithms.
Meanwhile, 70 percent said AI cannot be a true friend, and 22 percent said it can only simulate friendship. While one-third use AI for movie and music recommendations and a quarter for shopping, few would entrust it with traffic (8 percent) or financial decisions (5.5 percent).
Nearly half cited the loss of independence and critical thinking as the biggest risk of AI reliance, followed by 20 percent expressing concerns about privacy and data misuse.
