The rapid rise of generative artificial intelligence has intensified longstanding challenges in mathematics assessment, particularly for open-ended projects that can be easily outsourced to AI tools. Yet projects remain essential for fostering mathematical sensemaking, creativity, and positive mathematical identity, especially in liberal arts mathematics courses serving non-STEM majors.

This session presents a sequence of AI-resistant mini-projects implemented in a Liberal Arts Mathematics (MA111) course. These projects are intentionally designed around personalization, invention, and reflective explanation, features that require authentic student thinking and cannot be meaningfully completed by AI alone. Examples include: (1) designing an original recursive number sequence inspired by Fibonacci-type patterns, (2) analyzing mathematical structure in a natural phenomenon of the student’s choosing, and (3) creating a personal number system with defined symbols, rules, and representations.

The design framework integrates three core principles:

1. Personalization (student-chosen contexts and parameters),
2. Creation (students generate new mathematical objects or systems), and
3. Justification (written explanation of reasoning and meaning).

Student artifacts and reflections indicate that these projects promote ownership, creativity, and deeper engagement with mathematical ideas while substantially reducing AI-generated submissions. Practical assignment prompts, scaffolding strategies, and grading approaches will be shared so participants can adapt AI-resistant project structures to their own courses.