Heavy AI use sits alongside low confidence in the degree

The full distribution for “How often do you use generative AI tools (like ChatGPT) to assist or complete your academic work?” (n = 399) is: Daily / For almost all assignments — 57.9% (231); Occasionally for reference — 37.8% (151); Extremely rarely — 4% (16); Never — 0.3% (1).

The cohort is clearly concentrated on this question. A single option carries 57.9% of responses, so the headline figure is stable enough to plan against. The smallest recorded answer, never, still represents 1 people — a group large enough to matter in programme design even at 0.3%.

“Do you currently feel that your formal college degree can secure you an immediate, industry-ready job?” (n = 399) returns: Only partially sufficient — 45.1% (180); No, it is completely insufficient — 27.8% (111); Yes, it is fully sufficient — 27.1% (108).

Placed beside the primary figure, this measure spreads more evenly, with the leading answer at 45.1%. Read together, the two questions describe cross-cutting analysis as a pattern reported across the cohort rather than an issue confined to a minority of respondents.

“How do you primarily acquire modern, industry-relevant skills today?” (n = 399) adds: YouTube and free open-source materials — 50.1% (200); Strictly from my formal college classes — 32.8% (131); Paid online certification courses (Udemy, Coursera, etc.) — 15.3% (61); Government training setups (ASAP Kerala, K-DISC) — 1.8% (7).

This third figure is included because it constrains the interpretation of the first two: it shows where respondents locate responsibility or intent, and therefore which interventions the data can and cannot be used to argue for.

The sample is a community-recruited cohort of 399 respondents surveyed between 21 May and 28 June 2026. Figures describe this cohort; they are not state-level prevalence estimates and should not be reported as such.

The composition of the sample also shapes the totals: 74.4% of respondents fall in the older student and early-career band, and 72.7% identify as male. Any reading by gender, district or occupation needs disaggregated fieldwork this instrument was not designed to deliver.

The instrument records one moment in time, so nothing here identifies a cause or a trend. Statements about mechanism in this brief are interpretive and flagged as such.