ā Back to Library
Why self-reported data is a trap
Does Machiavellianism predict who cheats with AI? New research suggests the common assumption misses the mark entirely.
A study of 812 students investigated the link between dark triad traits and academic dishonesty. While many expect those high in Machiavellianism to exploit new technology for gain, the data showed no such correlation. Instead, the findings highlight a significant measurement problem where self-reported surveys fail to capture actual behaviour accurately.
This discrepancy is not evidence of restraint, but rather a flaw in how we conduct personality assessment. When participants under-report their actions and survey questions are poorly calibrated, the resulting data can mislead researchers and organisations alike. Understanding these limitations is essential to avoid drawing false conclusions about AI cheating in both academic and workplace settings.
Subscribe for weekly organisational psychology breakdowns, and let me know in the comments: do you trust self-reported survey data?
š Full research article: www.keca.co.uk/articles/who-gets-to-say-what-you-are
/ Official Transcript
Full Video Text
You'll hear that the calculating Machiavellian type is simply too careful to cheat using AI. That comes from one study of 812 students, where three dark traits predicted it and Machiavellianism did not. Now read what the authors said about their own result. Such people under-report and five of their questions failed.
It's not a finding about restraint, it's a measurement revealing what it couldn't see. What else are we counting this way? Okay. What else