Introduction:
Few would dispute that the implementation of artificial intelligence (AI) in education—both in general and specifically in entrepreneurship training—is now a reality. However, we are currently at a pivotal stage where critical analysis and strategic design are essential. From academic scholars reflecting on its implications to technology companies developing educational tools, key stakeholders are now facing crucial decisions. At this defining moment, it is imperative to turn our attention to both current and future users, understanding their fears and expectations to design AI-driven solutions that are truly adapted to their specific contexts and needs.
The objective of this study is to explore university students' trust in using AI in entrepreneurship education by analysing their beliefs, perceptions, particularly their fears, expectations, and overall sentiments towards its implementation.

Methodology:
The study adopts an exploratory and descriptive approach to understand university students’ perceptions and the level of trust generated using AI in their entrepreneurship training. The exploratory nature of the research allows for a deeper investigation into students’ attitudes, beliefs, and expectations regarding AI. By combining both quantitative and qualitative data, the study employs a mixed methods approach. This approach integrates Likert scale responses to measure the extent of AI usage in a structured, numerical format, along with open-ended responses to capture more nuanced, subjective views on the role of AI in entrepreneurship education. The use of mixed methods allows for a comprehensive understanding of the topic, combining statistical analysis with rich qualitative insights to offer a fuller picture of students’ trust on AI.
The data were analysed using SPSS (version 29) and Atlas.ti (version 23). First, the sociodemographic data and AI usage were analysed through descriptive statistics and histograms. For the qualitative responses, coding, thematic analysis, and word clouds were conducted. The final sample consists of 408 students.

Results:
A preliminary analysis of AI usage reveals slight differences by gender and age. Men report using AI slightly more than women; however, these differences are not statistically significant. Regarding age, a significant difference is observed at the 10% level (p < 0.10), with F = 0.244 and p = 0.622. Specifically, younger students (under 22 years old) use AI more frequently than those aged 23 or older.
Regarding the qualitative data analysis, the most frequently used student assessments, in order of occurrence, are essential, useful, complementary and important. Next, a word cloud was created along with a sentiment analysis to visualize the areas of trust and distrust among students regarding their entrepreneurship training with AI.

Conclusions:
The results show that the sense of trust in entrepreneurship training is stronger than distrust, particularly among younger students. However, many still express concerns regarding the efficiency and reliability of AI, highlighting areas where trust has yet to be fully established. By examining how age and individual expectations shape trust in AI, the study informs educators, policymakers, and AI developers on students’ needs and expectations. Ultimately, these findings contribute to the ongoing discussion about AI in education, highlighting the importance of building technology that students can trust.

Keywords: Artificial Intelligence, AI, trust, entrepreneurship, university, education, sentiment analysis.