The Learning Outcomes Problem: Why Traditional Assessment Fails in an AI-Assisted World

Here's the uncomfortable truth facing university administrators today: most of our assessment methods can no longer reliably measure what students actually know. When a large language model can produce a passable essay on 19th-century industrial economics or draft working code for a basic algorithm, the term paper and the coding assignment have lost their diagnostic power. This isn't a temporary glitch to wait out. It's a fundamental shift that demands we rethink what we're measuring and how we're measuring it. The good news? This crisis in assessment is also an opportunity. For too long, we've conflated the ability to produce written work with the ability to think critically. AI forces us to separate these competencies. The students who will thrive aren't necessarily the ones who can generate the best AI output—they're the ones who can evaluate that output critically, identify errors, refine prompts, and apply insights to novel problems. That's exactly the kind of higher-order thinking our institutions claim to prioritize. Practical steps start with faculty development - - not just training on AI tools, but deep conversations about learning objectives. What should a graduate of your chemistry program actually be able to do independently? Build assessments around those capabilities: live demonstrations, oral examinations, peer teaching, collaborative problem-solving with real-time reasoning. Consider portfolio-based assessment that captures process, revision, and reflection. The goal is to build mechanisms that measure what matters most when AI handles the routine work. Leadership here means creating space for experimentation. Some departments will move faster than others, and that's appropriate. Encourage pilot programs in assessment innovation, share results across colleges, and resist the temptation for top-down mandates that stifle faculty creativity. Your faculty are the experts in their disciplines. They'll find solutions that generic policy never could. But they need to be empowered to do so. The institutions that navigate this transition successfully will be those that treat their faculty as partners in reinvention, not obstacles to overcome.