018 - Detection Insights

Study Phase 3 Update: 50 individuals, over 550 pages of interview data, and over 5,100 segments of sequenced text have been gathered, formatted, and prepared for the next phase of the Cognitive Bias Detection System Study.

If all data is processed with precision certainty scores and all 188 cognitive biases are individually detected, this should produce roughly 960,000 data points for analysis. The goal of this analysis will be to isolate patterns of how cognitive biases are expressed over time by specific individuals.

11 frauds, 4 other criminals, 11 respected scientists, and 24 "Test" category individuals were selected. This includes interviews with the 2 most biased individuals (Tech CEOs) according to human bias detection scores from the study's first phase. Many interview records of famous criminals had been "lost" but using multiple search engines and archive.org a sufficient number of records were recovered to proceed.

If patterns specific to each group are successfully isolated, this could pave the way for early detection of both positive and negative patterns of bias, enabling prevention, intervention, and the cultivation of positive patterns.

Study Phases 1 and 2 published in pre-print earlier this month are currently going through peer review, slated for Open Access at Cell / Elsevier.

*To any other Cognitive Bias, Psychology, and AI researchers, I'd welcome co-authors for the Phase 3 analysis process and study paper. Nearly a million precision data points could take quite some time to analyze without assistance.

If successful, this method could be applied to the detection of many positive and negative patterns across every domain. After all, if you could detect both future Bernie Madoffs and Nobel Prize Winners through the same analysis method of an interview (or conversation) with them, what doors might that open? How much time and effort might that save?

*Note: See the pre-print of study phases 1 and 2, as well as links to the data here.

It is easy to imagine HR departments salivating at the thought of having a viable method of predicting the success of job applicants, Universities predicting academic success of potential students, financial firms predicting the risk of fraud, and governments predicting the risk of corruption. These capacities could become Dystopian if taken to extremes, but in moderation and sensible designs they could also have a more significant positive impact.