Computer science professor successfully used IBM's Watson to make an AI teaching assistant for answering questions online
Matt McFarland / Washington Post :
Context & Ripple Effects
Four years after its Jeopardy win, Watson had already been applied across 75 industries from finance to healthcare, and IBM was pushing it into new verticals on a weekly cadence — days earlier, researchers were training it to spot cybersecurity threats, and weeks later it would power digital ads that answer consumer questions. A professor wiring Watson up as a teaching assistant fits that same playbook: take the question-answering engine that beat Jeopardy and point it at a new corpus of domain knowledge.
First-order effects
- Students in the professor's course get online question-answering handled by Watson instead of by the instructor or TAs, freeing human time for higher-touch teaching work.
Second-order effects
- IBM gains another reference deployment for selling Watson into education alongside the ad-tech and security pushes already underway, each vertical case study feeding the next sales pitch.
Third-order effects
- If the pattern holds, domain-scoped Q&A assistants become a standard layer over institutional knowledge bases — a structure visible again in 2023 when IDG outlets launched Smart Answers, a chatbot trained only on their own publication archives.
The trend: Watson-era deployments mark the shift from one-off AI stunts toward embedding question-answering assistants directly into education, advertising, and publishing workflows.