Background: The LIRIK, an instrument for the assessment of child safety and risk, is designed to improve assessments by guiding professionals through a structured evaluation of relevant signs, risk factors, and protective factors. Objective: We aimed to assess the interrater agreement and the predictive validity of professionals' judgments made with the LIRIK in comparison to unstructured judgments. In study 1, professionals made safety and risk judgments for 12 vignettes with the LIRIK (group 1, n = 36) or without an instrument (group 2, n = 43). In study 2, we compared professionals' safety and risk judgments for 370 children made with the LIRIK (group 1, n = 278) or with no instrument (group 2, n = 92), with outcomes indicating actual unsafety in files 6 months later. Results: In study 1, agreement about safety and risks was poor to moderate in both groups. Differences between groups were small and inconsistent. In study 2, the predictive validity of judgments was weak to moderate in both groups. In neither group had unsafe outcomes increased consistently when unsafety or risks were assessed as higher. Conclusions: Judgments made with the LIRIK were not more reliable or valid than unstructured professional judgments. These findings raise important questions about the value of risk assessment instruments and about how professional safety and risk judgments can be improved.