Summary of qualitative themes and illustrative evidence
| Research question | Theme | Description | Illustrative quote | Interpretation |
|---|---|---|---|---|
| RQ3 | Task importance shaped checking | Participants checked less in low-stakes tasks such as brainstorming or drafting | “If it’s just an early draft or ideas, I don’t feel the need to check everything.” (S3) | Verification was often delayed rather than ignored |
| RQ3 | Confidence supported delayed checking | High AI literacy increased confidence in noticing errors later | “I trust it enough to move fast, but I also trust myself to catch mistakes if they show up.” (S6) | Literacy supported confidence, but not always immediate scrutiny |
| RQ3 | Effort–time trade-off | Participants weighed the value of checking against the effort required | “Checking everything defeats the purpose.” (S8) | Verification was shaped by efficiency considerations |
| RQ4 | Topic familiarity triggered doubt | Domain knowledge helped participants detect problems quickly | “When it touched something, I really know well, I stopped trusting it straight away.” (S9) | Background knowledge supported critical evaluation |
| RQ4 | Overconfident fluency triggered caution | Very polished answers to complex issues raised suspicion | “If it gives a very clean answer to a messy problem, that’s when alarms go off.” (S12) | Fluency sometimes prompted scrutiny rather than trust |
| RQ4 | Accountability triggered verification | High-stakes tasks increased checking behaviour | “If my name is on it, I check everything.” (S1) | Personal responsibility influenced active verification |
| Research question | Theme | Description | Illustrative quote | Interpretation |
|---|---|---|---|---|
| Task importance shaped checking | Participants checked less in low-stakes tasks such as brainstorming or drafting | “If it’s just an early draft or ideas, I don’t feel the need to check everything.” (S3) | Verification was often delayed rather than ignored | |
| Confidence supported delayed checking | High | “I trust it enough to move fast, but I also trust myself to catch mistakes if they show up.” (S6) | Literacy supported confidence, but not always immediate scrutiny | |
| Effort–time trade-off | Participants weighed the value of checking against the effort required | “Checking everything defeats the purpose.” (S8) | Verification was shaped by efficiency considerations | |
| Topic familiarity triggered doubt | Domain knowledge helped participants detect problems quickly | “When it touched something, I really know well, I stopped trusting it straight away.” (S9) | Background knowledge supported critical evaluation | |
| Overconfident fluency triggered caution | Very polished answers to complex issues raised suspicion | “If it gives a very clean answer to a messy problem, that’s when alarms go off.” (S12) | Fluency sometimes prompted scrutiny rather than trust | |
| Accountability triggered verification | High-stakes tasks increased checking behaviour | “If my name is on it, I check everything.” (S1) | Personal responsibility influenced active verification |