This paper investigates the influence of the Dunning–Kruger effect on workplace safety, focusing on incidents where overconfidence in less experienced employees may lead to safety breaches. Utilizing a backward analysis approach, we examine a series of incident reports to identify root causes and the role of human factors, particularly the cognitive biases associated with the Dunning–Kruger effect. Statistical inference will be employed using logistic regression to establish correlations between the perceived competence and actual safety outcomes. Our findings indicate that many safety incidents are deepened by someone overestimating their competency and understanding of safety rules. This work suggests insightful approaches to address this critical issue, including enhanced training programs focusing on cognitive bias awareness, adaptive testing methodologies better to assess competence levels and continuous performance feedback systems. Implementing these strategies aims to improve overall safety performance, reducing the likelihood of accidents caused by the misalignment of confidence and competence in the workplace.