A survey administered to Tulane University students in the fall semester revealed that, on average, students feel only moderately good about their knowledge of and competence using artificial intelligence tools.
Knud E. Berthelsen, a professor of practice in strategy, analytics and leadership minor and the Connolly Alexander Institute of Data Science, created the survey to gauge student attitudes, knowledge and usage of artificial intelligence at Tulane. The survey received 846 anonymous responses through a banner posted on Canvas.
Of the 846 respondents, undergraduate first-year and graduate students were the most represented in the sample.
To measure student AI competency, the survey asked respondents to rate on a scale of one through 10 how much they knew about AI, how well they understood how to use AI and how good they were at using AI. The total responses exhibited a relatively normal distribution, with many students rating their competency around five out of 10 across the three metrics.
These responses are consistent with students’ self-reported use patterns of AI tools. Unsurprisingly, most students had adopted AI tools, although with some exceptions: Slightly over 20% reported never having used AI. Another 20% of students report using AI daily, and a further 20% use it two to three times a week. The largest proportion of students reported using AI once a week in the fall semester.
These numbers roughly align with an international survey by the Digital Education Council, which was conducted in July 2024 and surveyed over 3,800 students from 16 countries. The survey reports that 86% of students regularly use AI for school, with 54% using it on a weekly basis.
Tulane students most frequently use ChatGPT, and are overwhelmingly more familiar with ChatGPT than other AI tools like Gemini, Copilot and Claude. The Digital Education Council reports similar data, with 66% of students saying they use ChatGPT.Students appeared to use AI mostly for academic support, for writing assistance and as an aid in research. However, students found that these tools were not well-suited to assist with complex mathematics and coding tasks, nor creative tasks like generating artwork. They also noted AI’s limitations in finding reliable sources and citations: the average rating for confidence in critically evaluating the output of AI tools was only slightly more than five out of 10.
Most students indicated that learning AI skills would be important for their future careers and should be implemented into Tulane’s curriculum. The respondents were required to indicate how much they agreed with the statement, with 10 indicating strong agreement and zero indicating strong disagreement. Approximately 66% responded with a five out of 10 or higher when asked if AI tools should be implemented more into Tulane’s curriculum. Of the 66%, 29% of respondents rated the importance of learning AI as a 10 out of 10. Similarly, 73% indicated a five out of 10 or higher when asked how important AI was for their future careers. Of the 73%, 24% of students responded with a 10 out of 10.
Despite the interest students express in learning about AI and how to more effectively incorporate it into their lives, all of the AI courses at Tulane during the spring 2025 semester had open space. DATA1250: Artificial Intelligence Tools was less than half full and DATA2810: Intro to Artificial Intelligence had only six enrolled students out of 25 by the end of the semester.
Berthelsen, through CAIDS, has been conducting a similar survey during the spring semester to assess changes in students’ attitudes towards AI. The results of the survey have not yet been released.