The single non-user: what saturation means for assessment

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 ai in education as a pattern reported across the cohort rather than an issue confined to a minority of respondents.

These are self-reported answers from a community-recruited cohort of 399 young Keralites surveyed between 21 May and 28 June 2026. They are not probability-sample estimates of Kerala's youth population, and confidence intervals from random sampling do not apply.

Because 74.4% of respondents sit in one age band and 72.7% identify as male, subgroup differences are indicative at best. Cross-tabulated claims require a larger, stratified sample.

Finally, the survey measures single-point self-report, so it can establish association and scale but not direction or cause. Where this brief notes a pattern across two questions, that pattern is descriptive.