Goal
Build an agentic digital twin of James Brusseau, professor of university courses in AI Ethics, and Business Ethics.
Instead of answering student questions as a generic computer, the agent mimics Brusseau as a particular person with unique vocabulary, style, and thought. (Note on mimetic AI here.)
Text and voice mimicking initially, visual avatar subsequently.
Technology adaptable to grade student work.
Technology adoptable for other professors and courses.
Milestones / Problems solved
Initial milestone: High fidelity office hours always open and judgment free ("The Caffeinated Professor never sleeps.")
- Student questions are learning opportunities, but professor unavailability collapses the potential into frustration. By responding on-demand, the agent converts uncertainties into immediate learning.
- Students can be anxious about asking questions face-to-face. Some fear being judged, others feel intimidated or awkward. Querying an agent may be psychologically liberating.
-
Students may have questions but remain uncertain about what, exactly, the question is. While human professors typically lack time for ranging discussions, exploratory exchange with the agent may surface the deeper question so that it can be addressed directly.
Derivative milestone: Grading assistant (Paradigm shift in student evaluation.)
- Grading student written work is tedious, a prime candidate for automation.
- Traditional grading reflects whether an answer is right or wrong. For Caffeinated Professor grading, all students reach the right answer. The question is posed, and students answer textually or orally. Then the agent provides cues toward improving the answer until a standard is reached. The subsequent grade reflects how much help - cues and nudges - the student required from the agent to achieve final understanding.
- Dynamic AI exams shift the grading paradigm from static learning interruption to dynamic learning contribution, while overcoming student reliance on AI for traditional written exams. (AI solves the problem of cheating and empty learning that AI itself caused.)