When undergraduate students are starting to learn their first programming language, they usually feel themselves overwhelmed by the sheer amount of information that appears in the IDE. Students are faced with different colors, a list of warnings, hints on how to deal with errors, and an output with a list of errors the compiler found in their code. However, even if students know which error to attend first, they need to clearly understand what the error is referring to. Unfortunately, newcomers to programming usually do not understand the error reports generated by compilers. As an attempt to help students to understand errors when programming, and to better learn how to program, we propose to use Generative AI tools. Students were asked to program a series of exercises. When they found an error, reported by the compiler, that they did not understand, they had to use a Generative AI tool that would explain the error to them. We tested this hypothesis with undergraduate students that were learning procedural programming for the first time. Analysis of the results showed no significant improvement of the test-group over the control-group. However, test-group students reported feeling more in control with programming than control-group students.
