Artificial Intelligence (AI) based assistants and agents are rapidly becoming part of how people write, code, study, and make decisions. While these tools offer unprecedented gains in productivity, they also introduce a growing risk: as people rely on AI assistants to generate ideas and solutions, they may gradually lose creative ownership and reduce their critical evaluation of suggestions. Over time, this shift can weaken independent reasoning and steer users toward conventional ideas rather than novel ones. As AI becomes deeply integrated into education and industry, it is essential that these systems be designed to strengthen, rather than erode, human creativity and critical thinking. This project addresses that challenge by investigating how people think when collaborating with AI assistants and by envisioning new forms of human-AI interaction that preserve and enhance human cognitive agency. The outcomes will guide the design of future AI systems that empower users to lead creative efforts and reason critically in an AI-augmented world. The project begins with empirical studies examining how users engage in critical and creative thinking while working with AI assistants on writing and programming tasks. Using established cognitive frameworks, the research team will analyze user behaviors, interaction patterns, and time allocation to identify where current AI interfaces create barriers or support deeper reasoning. These findings will inform the development of four novel interaction paradigms: structured deliberation that makes assumptions and evidence explicit, productive friction that introduces reflection at key moments, parallel co-thinking that supports simultaneous human and AI reasoning, and distributed cognitive roles that allocate subtasks transparently while preserving human oversight. To explore how these paradigms can be realized in practice, the project will develop and test prototype interfaces that examine alternative interaction and interface