This study examines whether awareness of implicit gender-science stereotypes influences university enrollment in STEM fields. We designed a randomized controlled trial involving 566 Italian high school seniors, combining surveys with an Implicit Association Test to measure unconscious biases. Before students finalized their university enrollment, a treatment group received personalized feedback on their IAT scores, while a control group received no information. Results show that revealing implicit stereotypes significantly reshapes educational choices, but with sharply contrasting gender effects. For women—who initially exhibited stronger stereotypes—feedback increased the probability of enrolling in STEM majors. Conversely, men with strong stereotypes who received feedback became less likely to choose STEM fields. These results highlight that awareness of implicit biases can be a powerful yet double-edged tool for addressing gender gaps in STEM education.