Enterprise AI chat logs: how Dartmouth's provost was caught
Dartmouth provost Santiago Schnell resigned after a faculty committee reviewed his enterprise Claude logs and emails and found undisclosed AI use in published work, including a Washington Post column and a ChemSystemsChem paper. The committee did not rely on Pangram or other AI detectors. The episode argues that the bigger story is the record your work AI account keeps, and debates whether patchy disclosure rules share the blame.
You come away knowing that Copilot, ChatGPT Enterprise and Claude prompts sit on employer servers, why logs are stronger evidence than detectors, and why you should…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.