This study investigated the phenomenon of cognitive surrender among students in the context of increasing AI tool usage in higher education settings, and examined the effectiveness of specific pedagogical activities designed to foster critical thinking and counter over-reliance on AI-generated outputs. The study was conducted across three universities in Uganda, involving 312 students from undergraduate programs in social sciences, information technology, and education faculties who were regular users of AI tools including ChatGPT, Google Bard, and similar large language model-based assistants. A mixed-methods quasi-experimental design was employed, with students in experimental groups exposed to structured pedagogical interventions over a 12-week semester period while control groups continued with conventional instructional approaches. Pre- and post-intervention assessments using the WatsonGlaser Critical Thinking Appraisal (WGCTA) were administered to measure critical thinking competency changes. Data were also collected through questionnaires, focus group discussions, and learning portfolio analysis. Findings indicated that cognitive surrender, defined as the uncritical acceptance and wholesale adoption of AI-generated content without independent verification, evaluation, or intellectual engagement, was prevalent among 67.3% of surveyed students prior to intervention. Post-intervention assessment showed that students in experimental groups demonstrated significant improvements in critical thinking scores (mean gain = 8.4 points on WGCTA) compared to control groups (mean gain = 1.9 points), confirming that targeted pedagogical interventions effectively promoted critical thinking and reduced cognitive surrender behaviors. The most effective pedagogical activities were structured AI output critique exercises, Socratic questioning seminars, collaborative fact-checking workshops, and AI-human comparative analysis tasks. The study concluded that proactive pedagogical design was essential to preserving critical intellectual agency in AI-augmented learning environments and recommended the integration of AI-critical pedagogy into curriculum frameworks across Ugandan higher education institutions.