Can On-Device LLMs Do Private Qualitative Coding?
Qualitati Research Team · 2026-06-01 · 6 min read
Last updated: June 1, 2026
Short answer
On-device, open-source LLMs make private qualitative coding technically feasible. In a 2026 CHI study introducing ChatQDA (Ngo, Nguyen Van, Nguyen, Do & Nguyen-Quoc, 2026), four researchers used the open-source gpt-oss-20b model locally to code an interview transcript. They showed "conditional trust" — valuing AI for basic extraction (M=3.75 on speed) but doubting its interpretation, and they rated perceived security lowest (M=2.75). The lesson: privacy must be verifiable, not just promised, and interpretation stays human.
Where Qualitati fits
Qualitati's ThemeLens proposes themes anchored to participant quotes, and the QDA Workspace lets researchers review and validate every AI-assigned code before it becomes a finding — extraction speed with human-in-the-loop interpretation. See also our guide to LLM inter-rater reliability. View transparent pricing.
Full article available at qualitati.com/blog/on-device-llm-private-qualitative-coding-2026.