Chat GPT is really of limited use, at least in my discipline (history). Just to play around with it, I’ve asked it to generate sources for bibliographies – it has given me lists of books that are tangentially related to the topic, often overlapping lists for different questions, and not entirely the kinds of sources I was looking for (primary sources when I was asking for secondary, etc.). Same with outlines and such: generic, not original, not the product of actual research, not tailored to research questions.
Bottom line: using Chat GPT gets you out of doing the hard work of thinking through and organizing a paper – but not terribly well. Is this useful? I’m sure the tool will get more sophisticated, but right now, students are better off skipping it.
As far as a general culture of academic integrity is concerned, there may be some schools where such a culture exists, but generally, it comes down to individual departments and professors establishing policies and strategies to keep cheating at a minimum. How do they design research projects? Are in-class exams making a comeback? Do faculty use plagiarism detection (and AI detection)? Is there a cutthroat student culture (college-wide or at the departmental level) that would promote cheating even among high-achieving students? I think the answers to these questions will vary across departments and individual faculty at most universities, regardless of school culture.