Poll: Majority of Higher Ed Leaders Optimistic About Generative AI

Everyone’s Talking About Generative AI in Higher Education

For a tool that debuted as recently as late November, ChatGPT has done a remarkable job of penetrating the academic consciousness.

Only 21 respondents in total had no familiarity with ChatGPT or generative AI before the survey was sent to them, and 46 percent of respondents said they were either very or extremely familiar with it, a level of familiarity that was consistent among the survey’s four constituent groups: C-level administrators, directors/managers, staff and faculty.

Perhaps unsurprisingly, however, one of those groups had a harsher view: 28 percent of faculty had a pessimistic view of generative AI, more than twice the overall rate, although that group also had the lowest number of neutral responses (21 percent). Nearly half of all faculty members surveyed (49 percent) still said they were optimistic or very optimistic. The other cohorts shared similar rates of optimism, between 54 and 57 percent.

READ MORE: Here are three trends shaping modern higher ed classrooms.

The survey also showed that conversations about generative AI aren’t just limited to social media — where, to be fair, there have been plenty. College and university leaders are already engaged in conversations among themselves, with 75 percent saying the topic has been discussed somewhat, quite a bit or a great deal. Those conversations figure to continue as the AI improves and administrators become more aware of its capabilities.

As for what those conversations are leading to, respondents identified a number of areas where generative AI, including ChatGPT, has already made an impact. Fifty-four percent identified at least one of 13 possible areas of impact offered by the survey’s authors; all of the most common areas were related to instruction: undergraduate teaching (37 percent), teaching and learning centers (32 percent), instructional technology (31 percent), graduate teaching (30 percent) and faculty development (30 percent).

Interestingly, respondents also were asked which areas are most likely to be affected in the future, and the most popular response to that question had nothing to do with instruction. The top two responses: 51 percent said faculty research would be impacted by generative AI sometime in the future, and 52 percent said IT was likely to be impacted.

Given an opportunity to comment further, respondents shared an array of reactions by their universities to ChatGPT, with one saying the institution had already conducted six studies on integrating generative AI. A faculty respondent added that they were already using the tool to “develop course topics.”

How Is Generative AI Being Used in Higher Education?

While there is a great deal of interest and discussion around generative AI on college campuses, the survey also found that there is still plenty that administrators and instructors don’t know about how the tool is being used.

When pressed to answer how students were currently using generative AI, “I don’t know” was the overwhelming choice at 55 percent, nearly double any other answer.

As for what was once the biggest question about ChatGPT, 23 percent of respondents thought students were using the tool for “submitting generated material without editing it,” what most institutions would consider a straightforward form of plagiarism. Slightly more (29 percent) believed students were submitting generated material after editing it. Others believed students were using it for benign or even educational purposes, with 27 percent of respondents believing it was being used both for ideating or for entertainment. (Respondents were allowed to select more than one answer to this prompt.)

LEARN MORE: How universities can use AI chatbots to connect with students.

The survey also asked how faculty and staff were using generative AI for student learning activities, and while the runaway winner was again “I don’t know” (53 percent), a portion of faculty and staff were already using it for things like in-class activities (24 percent), discussion prompts (22 percent) and homework and course assignments (22 percent).

Faculty were also using generative AI for their own purposes, with 24 percent using it to write emails, 23 percent using it to write code and 21 percent using it for creating instructional content and writing reports.

How Concerned Is Higher Ed About Cheating Using Generative AI?

Despite the overall optimism for the tool and its potential positive uses, respondents to the survey still were most concerned about student cheating. When asked their greatest concerns relative to generative AI, the top response was academic integrity (75 percent), followed closely by two answers both reflecting the ways in which generative AI still falls short.

Sixty-eight percent of respondents said they were concerned about over-reliance on outputs and inaccurate outputs, both of which are issues that those trying to tamp down the early panic over ChatGPT pointed out. While the GPT 3 is good, it is in most cases still not good enough to fool a college professor or complete a course with the level of rigor expected at most higher education institutions.

Other ethical concerns were cited, with 58 percent worried about data privacy and security, and 53 percent concerned about issues of diversity, equity and inclusion, which follows broader concerns about bias in AI.

On the other end of the spectrum, respondents were most positive about generative AI’s potential for improved efficiency (77 percent) and upgrades to existing personalized or adaptive learning tools (59 percent).

The full survey results and EDUCAUSE’s analysis can be found here.

UP NEXT: Microsoft makes a major investment in ChatGPT tech.