Generative Artificial Intelligence (AI) tools (e.g., ChatGPT) are rapidly reshaping the educational and professional development of STEM students. While prior work has largely focused on the technical capabilities of such kinds of AI tools, less attention has been given to the peopleware dimension, including behaviors, perceptions, and cognitive implications of students when dealing with them. This short paper explores the experiences and perspectives of STEM students enrolled in Mathematics and Computer Science curricula about the use of generative AI tools in current coursework and in near future. We conducted 12 semi-structured interviews and analyzed the data using template analysis. Our preliminary findings indicate that students perceive these AI tools as valuable for facilitating understanding and optimizing work, but they also report concerns related to cognitive laziness, reduced critical thinking and creativity, and over-reliance and anthropomorphization. Students broadly expect AI to replace humans in specific tasks and reported adapting their educational strategies through the acquisition of AI skills and increased specialization. These findings highlight emerging cognitive and professional trade-offs associated with AI adoption and provide an initial baseline to inform educational strategies that balance AI-assisted productivity with the preservation of human cognitive engagement and creativity.