“What Am I Really Doing with My Own Brain?”: College Students on Learning, Ownership, and AI

Synthesis: Ting, Liang, Tankersley, Lindgren, Ingram, Dallavalle, Lund and Liang interviewed 14 undergraduates at a Northeastern liberal arts college in summer 2025 about where they draw the line between using generative AI to support their thinking and letting it replace thinking they believe they should still be doing themselves. The line turned out not to be about use versus avoidance: students accepted AI for organizing, translating, summarizing, generating practice material and giving Feedback, and drew it instead at having the tool produce work they felt they had not done. Their reasons for holding that line were heterogeneous — developing skills, thinking independently, judgment they could trust without the tool, ownership and pride, and moral concern — and the authors argue these reasons should not be collapsed into one motive. Purpose entered fewer decisions than expected: some students connected an AI choice to a purpose or future professional identity, while others with a clear purpose made no such connection. What students valued most consistently was staying the one directing and judging the work — a distinction the paper aligns with dependent versus autonomous offloading — but deadlines, workload, grades and peer comparison regularly overrode it.

Key Findings

1. The boundary is about what gets handed over, not whether AI is used. Students described AI as acceptable for organizing, explaining, translating, summarizing and critiquing their own work, and unacceptable when it produced the work itself: "use AI to enhance, not to create."
2. Their reasons for keeping work their own are distinct motives. Worries about losing critical-thinking and writing skill differ from wanting to feel pride in authorship, which differs again from moral objections about environmental cost, plagiarism and originality; the authors argue research should examine these separately.
3. Purpose explains less than the field assumes. Of 14 participants, one directly tied an AI refusal to her purpose, one worried AI would erode skills her purpose depends on, three had future professional goals shape their thinking, and several others described purposes without connecting them to any AI choice.
4. The same student can reason both ways. Participant 2 said the grade should reflect her own work and then, under deadline pressure, that "there's no time to hold on to my pride right now. I need to just use AI to get it done."
5. Task interest shifts use within a person. Participant 7 wanted his own thoughts in academic work but said of an uninteresting field, "I would just let AI do it," which the authors read as evidence against a fixed per-student AI-use style.
6. Students describe the same worry their instructors have. They asked where AI helped and where it substituted for something they needed to learn, and the paper argues the student–faculty conversation is framed too narrowly around misconduct when students themselves name learning loss.
7. Conditions, not only character, shape the choice. Workload, deadlines, grades, convenience and fear of falling behind peers who use AI all pushed toward greater reliance, so the authors argue a response confined to changing student behavior will miss the setting that produced it.

Where the Line Falls

The distinction students drew most consistently was between supporting a process they were already running and turning over the thinking itself. Participant 8 stated it cleanly: "If it's like I've already done work and then this is helping me organize that work, then I think that's what I would view as okay. But if I'm having it produce work itself, that's where I draw the line." For Participant 1 an AI answer was the starting point rather than the product — "where my own thinking starts" — after which she cross-checked it and built her own response. One international student described using AI to express ideas he already had in English and said, "I personally feel like it enhances my voice," which is the mirror image of the loss-of-authorship story: the same tool can carry an idea the student already holds.

The authors treat this as a version of the dependent versus autonomous offloading distinction, and their addition is the student-side account of it. Offloading did not equal losing Learner Agency in these interviews, because several students were treating AI as a subordinate instrument they directed and judged. What separated them was whether the student remained the one deciding what the work should say — a metacognitive condition, not a property of the tool, and one no usage count would detect.

Distinct Reasons, Not One Motive

Students' reasons for doing parts of the work themselves did not reduce to a single concern about learning. Some worried about competence: Participant 1 said AI made work quicker but that she was "losing the ability to do it on my own," and Participant 5 asked the question the paper takes as its title. Others cared about ownership and pride: Participant 10 wanted to know which thoughts were actually his and to "take a little bit of pride" in work that was his own. Participant 12's reasons were moral — environmental harm, plagiarism, originality and creative ownership — and Participant 3 simply felt wrong submitting work that was no longer fully hers.

The authors make a methodological argument out of this. Skill erosion, felt authorship, and ethical objection arrive through different mechanisms and would respond to different interventions, so a study that pools them into "concerns about AI" measures an average of things that behave differently.

Purpose Enters Unevenly

The paper's sharpest negative result concerns purpose, defined as a personally meaningful intention to contribute beyond the self rather than any career goal. Participant 12 connected hers directly, saying of AI's costs that she "can't use it if I want to maintain and preserve my sense of purpose." Participant 1 worried AI was weakening "skills that I value and skills that I hope to use in whatever my purpose is." Participants 3, 5 and 13 let future professional identity shape their choices — "doing the work contributes to the education and thus contributes to the psychologist that I will be out in the world one day."

Others did not connect the two at all. Participant 2 located purpose in relationships and said AI "does not affect that"; Participant 8 held a strong boundary against AI-produced work and a clear vocation toward social work without linking them; Participant 10 described his purpose as "bringing joy to others" and called purpose and AI "kind of separate." The conclusion is deliberately modest: having a purpose does not mean drawing on it when deciding how to use AI, and the paper names the question of why some students make that connection as open. That matters for anyone hoping that purpose-driven framing — selling the purpose — will govern behavior on its own.

When Pressure Beats the Value

Students' stated values did not always survive the conditions they were studying in. Participant 2's reversal is the clearest case in the paper, and Participant 13 called AI under time pressure "that devil on your shoulder" and "the easy way out." Participant 3 began using AI partly from fear of falling behind classmates, and Participant 5 found AI summaries attractive when "swamped during finals." Interest modulated the same pressure: Participant 7 let AI take uninterested material while using it to go deeper in fields he cared about.

The authors draw an institutional rather than an individual conclusion. If grades weight the end product over how students think and grow along the way, pressure toward AI use follows even for students who would rather do the learning — so a response confined to student behavior will miss the setting that produced it. They propose making the reasoning visible: early ideas, decisions along the way, revisions, and an explanation of what changed and why, with process evidence valued over the artifact. They are explicit that this is not a detection strategy but a statement about what the assignment is for.

What this means for practice

• Instructors. Name, in the assignment itself, which parts of the thinking students must do themselves and why those parts matter, instead of only stating what AI use is permitted. The interviews show students already asking that question and guessing at the answer from course-level rules that varied by instructor.
• Instructors. Sequence the own-attempt before the AI turn in at least one high-stakes task: students first work the problem or build the argument, then use AI for feedback, comparison or revision, so the tool lands after the effort instead of before it.
• Assessment designers. Grade evidence of how the thinking developed — early ideas, decisions, revisions, an explanation of what changed — so that doing the work remains worthwhile when AI can produce a better-looking product faster.
• Administrators. Treat workload, grading weight and deadline structure as AI-policy variables. Students in this study named them as the reason their own values lost, and no amount of guidance will outrun a course that makes reliance rational.
• Researchers. Do not treat a stated purpose as evidence that purpose governed an AI decision. Code the connection explicitly, as this study did across four categories, and expect a large share of participants to fall outside any purpose–AI link.

Limitations

• Fourteen undergraduates at one Northeastern liberal arts college, recruited by convenience and snowball sampling, so the accounts reflect one institutional setting and cannot represent college students generally.
• The sample skews toward psychology, applied psychology, neuroscience and related majors, and most participants reporting class year were juniors or seniors, which may have shaped which concerns surfaced.
• The interviews were not recruited iteratively and the authors explicitly decline to claim thematic or meaning saturation, particularly for purpose; the protocol was designed for a broader study of AI experiences, so 35–45 minute interviews constrained depth on any one topic, and participants varied in what they understood "purpose" to mean.
• All accounts are retrospective self-reports, so consistency between stated values and actual behavior across situations cannot be established, and no causal claim about reasons, circumstances and AI-use patterns is supported.
• Nothing links students' worries about losing skills to measured change in their writing, critical thinking or other abilities; the paper states this plainly rather than implying the concerns are calibrated.
• One of the undergraduate interviewers later joined the analysis, and the authors disclose that his own experience may have made him more likely to notice purpose–AI connections, describing the counterexample checks they used in response.

Citation

Ting, T., Liang, E., Tankersley, M. E., Lindgren, T., Ingram, N., Dallavalle, C., et al. (2026). "What Am I Really Doing with My Own Brain?": College students on learning, ownership, and AI. Journal of Intelligence, 14(11), 273.