Generative AI & Academic Anxiety
Mapping how student cohorts deal with moral ambiguities, pressures to adopt technologies, and responses to perceived increased surveillance.
Key Quantitative Metrics
AI Resisters
* This was a direct result of the LCA modeling, refer to the Data Analysis section.
* This was a result of the qualitative analysis of the open-ended questions.
The Student Experience Spectrum
Latent Class Analysis (LCA) modeling identified three student groups across campus.
- Moral Injury: They may experience stress from playing by the rules while others don't.
- Diminishing Returns: Their high effort no longer correlates in an obvious way with academic reward.
- Anxiety of Detection: They may have constant underlying fear of being accused of using GenAI.
- The Arms Race: They may feel forced to use GenAI to compete with peers who are using it, even when such use comes into tension with their personal ethics.
- Skills Mismatch: They may feel that their classes are teaching 'outdated' skills that are no longer useful in the workforce.
- Value Disconnection: They may disengage from tasks perceived as "busywork" and that are easily automated.
In-Depth Implications for Wellesley
Expand each section below to read the qualitative findings from the complete report.
As GenAI becomes increasingly ubiquitous within institutions of higher education, students are experiencing both technical and moral uncertainty as they attempt to navigate the unprecedented technology and its effects. What has emerged from the data is that through this uncertainty, many of our students are developing their own moral frameworks for GenAI use rather than simply following the guidance of their professors or the administration.
The mismatch between the speed of the introduction of emerging technologies and our ability to understand how they work and how they are changing our world can lead to feelings of uncertainty about the present, the future, and how we ought to respond. Moral uncertainty is a state of being that arises when people face crucial decision-making without knowing where to turn for moral guidance. 52 participants expressed moral uncertainty, in which they do not know how they feel about its ethics. While 69 participants are uncertain about a world with advanced AI.
"Everyone's using it but we're not getting any information on how to use it properly/ethically" (Participant 33)
Some students expressed an internal moral friction as they wrestled with their desire to refuse GenAI tools because of ethical concerns, while at the same time feeling like GenAI literacy will be important for their future career:
"[I am] anxious because I worry about having to go against my values and beliefs in order to attain a job that I want, even if it doesn't use AI heavily" (Participant 147)
At the same time, some students expressed minimal moral uncertainty, turning instead to absolute stances against cutting corners on human tasks or outlining the vast environmental downsides of the technology, noting its impact on disadvantaged communities.
Our data shows that our highest category of students could be categorized as AI Resistant at 43%. While previous research has shown that professors distrust their students and assume that they are using GenAI, our survey results suggest that more students refuse to engage or are resistant to adopting GenAI into their academic careers. 66 participants who refuse to use AI framed this decision as an ethical rather than technological one.
Participants demonstrated awareness of the consequences of their use, and resisted complicity in the harms of AI. Environmental concerns and water waste were oft-mentioned concerns.
"I wish more people were discussing the environmental impacts of generative AI, particularly on communities where the water quality is being impacted" (Participant 215)
As a result, 38 participants expressed discomfort when given assignments made by or requiring the use of GenAI. When policy requires AI use for assignments, many indicated they prefer not to follow it, stating a deep desire that required automation should not become standard across coursework portfolios.
Despite a lack of expertise on GenAI and its educational implications, some students appear to be exercising individual agency within their courses and assignments. Faced with a perceived pressure to be AI literate to increase employability and uncertainty about how AI will affect their careers, students have resorted to using their own judgment and decision-making to decide how to use AI in courses as a training ground for work, regardless of AI policies.
30 participants indicated a desire for technical literacy, claiming that they want to learn how to use GenAI as a tool that will aid them in their future careers. 12 participants specifically used the term “under” or “unprepared” when describing how they feel about GenAI in the workforce.
"In today's world it is a disadvantage to not be able to know how to manipulate AI into helping and bettering our work. We need to learn how to use it in a way to make us stronger, but not rely on it to replace our thoughts" (Participant 46)
62 participants demand workforce positioning and career adoption, in which students learn how GenAI is impacting their desired career and plan their education to accommodate that. 30 participants expressed fear that sheltering students from GenAI as a result of classroom bans and lack of AI education cuts them off from potential career prospects.
Following repeated instances of prohibited AI use, many institutions and professors have started relying heavily on AI detection platforms. As of January 2026, Wellesley College has made the GenAI-detector software Pangram available to all faculty. Since its adoption on campus, some students have developed negative feelings associated with the perception of having their work and their behavior subject to surveillance.
Specifically, 63 participants commented on the fear of surveillance and false detections due to the presence of GenAI and the mistrust it perpetuates. This climate of suspicion has fundamentally shaped the way that students go about writing their assignments.
Overall, 60% of our survey participants have changed their writing style to avoid detection or the threat of false accusations before submitting assignments. Students are avoiding em dashes, altering their vocabulary, and changing sentence structures to avoid having their writing mistaken for text generated by GenAI, even when they didn’t use it to write their assignments.
"I wish they understood that AI detectors can be wrong quite easily; it's really awkward to be accused of using AI and have no way to prove you didn't besides just your word" (Participant 216)
Co-Designed Recommendations for Students
The proposed recommendations for students were written by our research team, and the paragraph beneath each recommendation was paraphrased based on the results of the focus group (refer to the section "Focus Group" in the Methodology page).
Have informal discussions with your peers who may think differently than you about the role of GenAI on our campus.
Having conversations with peers about GenAI outside of the classroom can allow you insight into how other students, especially those who may come from different backgrounds or have different opinions on GenAI, think. Having these discussions will allow you to develop a more nuanced understanding of the technology and prepare you to develop or defend your opinion.
Develop critical GenAI literacy on your own, using online resources, tutorials, etc.
Learning how to use different platforms such as Gemini, Claude, and ChatGPT can teach you how to understand how GenAI works. Understanding what data these platforms use and how they generate answers will allow you to understand what their strengths and weaknesses are, and how to avoid hallucinations or misinformation. Having these skills might also help you to spot GenAI when it is unclear or not marked as such.
Think carefully on your own about how to avoid letting GenAI replace important skills that require work and habitual practice, like carefully reading a difficult text, translating one language into another, etc.
The AI Resisters in the focus group particularly responded to this recommendation, arguing that it is important to reflect on how certain skills or aspects of your education have changed since GenAI has become more popular. If you recognize that you are relying on GenAI for certain skills that you were once able to do yourself, think about ways to resist using it as a crutch.
Do your own research about how GenAI may impact your desired field of study or industry of work.
Reach out to people working in your desired industry and ask them how they experience GenAI day-to-day. Hearing personal stories about how professionals are being asked to utilize or avoid GenAI in the workplace, as well as if GenAI has replaced any jobs in their company, could provide students with more insight into how GenAI is impacting their career outside of the statistics they read in headlines.
Create or participate in a student-led advisory board or organization that offers a forum for students to discuss GenAI, including but not limited to the moral and political dimensions of the tool, its appropriate role on campus, etc.
Both the AI Resisters and AI Ambivalent members of the focus group agreed that having a student-led organization would allow students to have a space to both discuss the way they feel about GenAI and also learn more about the topic to develop their own perspective.
The AI Ambivalent members noted here that one who is ambivalent about GenAI isn't necessarily neutral or unopinionated, but rather that their opinion is complex and nuanced, and not as black and white as resister versus adopter.
Create or participate in a student-led organization that offers a space for students to learn about how to use GenAI tools from one another.
Both the AI Adopters and AI Ambivalent members of the focus group agreed that this is an important recommendation to implement in a low-stakes environment so that students can focus on learning how to use GenAI outside of the classroom.
Learning when GenAI use is appropriate, and how to use it properly, is an important skill for a curious student who does not want to engage with the technology improperly.
Decide on your own where to turn for moral guidance on appropriate use of GenAI amidst uncertainty.
Students can turn to family members, peers, professors, or authors and researchers to guide them as they think about the impacts of GenAI and how they want to engage with the technology.
Request clarity or further explanation from a professor when you perceive friction or a lack of understanding with respect to your professor's AI policy.
Politely ask your professor to provide you with examples of when AI use is allowed and when it is prohibited if the policy is unclear. If you have a question that would be relevant for the entire class, asking for clarification while the professor is discussing the syllabus is appropriate. If you have specific questions about the policy, such as if a particular website or use is permitted, book a time during office hours for further clarification.
The AI Ambivalent members were particularly drawn to this recommendation because if one does not have their own opinion on GenAI or has a complex opinion, then it is easier for them to follow their professor’s policies.
Advocate either on your own or in a group for Wellesley College to offer courses that critically engage with GenAI as a moral, political, and labor tool.
Interestingly, both our AI Adopters and Resisters stated that this was an important recommendation for students: having an organization or platform to discuss GenAI, whether it is something you embrace or reject, is important for developing a nuanced perspective on the topic.
Advocate either on your own or in a group for Wellesley College to offer courses that teach students how to use GenAI.
Ask departments that are relevant to the AI discussion, such as Computer Science, Philosophy, and Cinema and Media Studies, if there are classes they recommend that deal with conversations about GenAI.