Critique: AI is making college students change majors

Source: axios
Authors: Avery Lotz
URL: https://www.axios.com/2026/04/02/ai-college-students-change-majors-poll

What the article reports

A 579-word Axios brief reports Lumina Foundation-Gallup polling (Oct. 2025, n=3,801) showing that nearly half of enrolled college students have considered changing their major because of AI's potential job-market impact, and 16% already have. The piece includes demographic breakdowns by gender and field of study, notes a gap between student AI use and institutional policy, and quotes two sources — a student and a Lumina Foundation executive.

Factual accuracy — Adequate

The polling figures are sourced specifically to Lumina Foundation and Gallup, the fieldwork dates (Oct. 2–31, 2025) and sample size (3,801) are disclosed in the methodology note, and the demographic breakdowns (e.g., "male students (60%) … female students (38%)") appear internally consistent. One copyediting error is present — "asked them asked about" — but it does not affect a factual claim. The Christina Eid survey statistics (12% in 2024, 30% in 2025) are attributed to her own AU Kogod School survey, which is a narrow, single-institution, self-administered poll; the article does not note this limitation. No verifiable figure appears outright wrong, but the Eid data is not independently checkable as presented.

Framing — Promotional

1. Headline as settled fact: "AI is making college students change majors" presents a causal claim the polling only supports correlatively — the survey measures consideration of changing, not documented AI causation. The headline overstates what the data show.
2. Unattributed cause-effect: "AI isn't just impacting how academics study — it's also determining the studies they pursue" is authorial voice, not a quote or finding. The polling shows association, not determination.
3. "Threat level" section header: Axios's house style uses this label, but applying it to the finding that colleges discourage AI use implies colleges are the threat — framing an institutional-policy debate as a danger narrative without attribution.
4. Closing advice as editorial: Eid's "Get on with it" closes the piece as a bottom-line instruction. The article does not balance this with any voice skeptical of rapid AI adoption, lending the piece a pro-AI-adoption tilt in its final impression.

Source balance

Ratio: 2 human sources, both broadly supportive of AI engagement. Zero voices from: faculty skeptical of AI, students who chose not to change majors because of AI concerns, higher-ed researchers studying labor-market outcomes, or employer perspectives. The poll's sponsor (Lumina Foundation) is also the only institutional human voice quoted — a potential conflict of interest that goes unacknowledged.

Omissions

1. Prior-cycle context: Have students historically changed majors in response to other labor-market disruptions (automation, 2008 recession, offshoring)? That base rate would help readers gauge whether 16% is unusually high.
2. Actual labor-market outcome data: The article asserts graduates face a "rocked" job market but cites no employment statistics. Readers have no way to assess whether AI has measurably reduced entry-level hiring.
3. Skeptical or critical voices: No faculty, higher-ed researcher, or student who resisted changing majors is quoted. The strongest concern raised — Courtney Brown's worry about bias literacy — still presupposes AI adoption is coming.
4. Lumina Foundation's institutional interests: Lumina funds college-access initiatives and commissioned this poll. That context is absent, leaving readers unaware that the sponsor has a stake in narratives about college relevance and adaptability.
5. Margin of error / confidence intervals: The methodology note gives sample size and dates but no margin of error, which matters when comparing subgroup figures (e.g., 52% engineering vs. 54% business — a difference potentially within noise).

What it does well

• Demographic granularity: The breakdowns by gender and field of study ("vocational (26%) and tech (25%) majors") give readers a richer picture than a single topline figure would.
• Policy-practice gap: The finding that students use AI daily even "at schools where it is prohibited" ("10% use it daily and 17% weekly") is genuinely newsworthy and concisely reported.
• Methodology disclosure: Fieldwork window, sample size, and rounding note are included — more than many news briefs provide.
• Quote specificity: Eid's year-over-year figures ("12% of students … 30%") give a concrete data point rather than vague anecdote, even if the survey's scope is limited.

Rating

Overall: 6/10 — Solid data journalism scaffolding undermined by one-sided sourcing, a headline that overstates causation, and omission of context that would complicate the adoption-is-inevitable narrative.