TOPICS: HOW AI SPEAKS in Spring 2027 (LI1910A)

Large language models speak, translate, explain, and invent with startling fluency. But when can we tell human language from machine language, and when can't we? This Digital Research Lab treats that as an open scientific question. You'll build your own evidence, recording your speech, gathering your writing, and generating matched machine samples, then defining linguistic features precisely enough for software to count. You'll work at two scales, closely reading a single pair of texts and counting across a pooled class corpus, and learn which claims each can support. One strand turns to fiction, setting your stories against machine-written ones from the same premise. The course also takes up privacy, consent, and fluently fabricated citations. Any language welcome. No prior linguistics, statistics, or programming.