The rapid rise of generative artificial intelligence (GenAI) tools is transforming the higher education landscape, challenging the validity of long-standing assessment practices. In engineering education, where mastery is often shown through written reports and problem-solving, these technologies raise important concerns about authenticity, authorship, and skill development. This study explores the potential of student-created video presentation coursework as an AI-resistant alternative to traditional assessments. Drawing on a case illustration from a final-year engineering module, qualitative student feedback was analyzed to identify perceived benefits, challenges, and features that enhance authenticity and learning impact. The findings highlight video presentations as a powerful vehicle for fostering critical thinking, communication skills, self-reflection, and personal ownership of work, while naturally limiting opportunities for AI-driven academic misconduct through their embodied and performative nature. The study distills these insights into practical design implications for educators seeking to implement process-oriented, multimodal assessments that align with the demands of an AI-rich learning environment. In doing so, it contributes to the growing discourse on assessment innovation, offering evidence-based guidance for sustaining integrity and meaningful learning in the era of GenAI.