Recent research argues for a formalized hybrid risk assessment model that combines the current online child sex abuse risk measures with digital forensics artifacts. We conducted a feasibility study as an initial step toward formalizing the hybrid risk assessment model by identifying high-level digital forensic artifacts that have the potential to be valid and reliable indicators of risk, with a focus on CPORT Items 5, 6, and 7. Law enforcement investigators from a High Tech Crime Unit (HTCU) randomly selected seven closed cases; selection criteria included: male offender over 18, mobile device, child sexual abuse material (CSAM) offense, and 2019–2023 index offense. Investigation details related to probable cause, final charges, conviction, and offender risk were not disclosed. Statistical information ( %) for the following digital forensics artifacts was examined: 1) pornography collection (e.g., % of media, content type, gender ratio) and 2) evidence of networking/grooming and other problematic online activities (e.g., number of native messages vs. application messages; type of installed apps). The analysis predicted whether the offender was a CSAM-only or dual offender and if our findings agreed with the level of risk for reoffending suggested by CPORT Items 5, 6, and 7. Results were shared with the HTCU and scored for accuracy. The hybrid model was accurate in 6 of 7 cases. We conclude a hybrid model is feasible, and the findings illustrate the importance of analyzing app artifacts for context. Study limitations and future research recommendations are discussed.

中文翻译：

---

测试混合风险评估模型：根据数字取证工件预测 CSAM 犯罪者风险


最近的研究主张建立一种正式的混合风险评估模型，将当前的在线儿童性虐待风险措施与数字取证工件结合起来。我们进行了可行性研究，作为正式化混合风险评估模型的第一步，通过识别有可能成为有效且可靠的风险指标的高级数字取证工件，重点关注 CPORT 项目 5、6 和 7。高科技犯罪部门 (HTCU) 的执法调查人员随机挑选了 7 个已结案的案件；选择标准包括：18 岁以上男性犯罪者、移动设备、儿童性虐待材料 (CSAM) 犯罪以及 2019-2023 年指数犯罪。有关可能原因、最终指控、定罪和犯罪风险的调查细节尚未披露。检查了以下数字取证工件的统计信息（%）：1）色情作品收集（例如，媒体百分比、内容类型、性别比例）和 2）网络/诱骗和其他有问题的在线活动的证据（例如，本地人的数量）消息与应用程序消息；已安装应用程序的类型）。分析预测犯罪者是仅 CSAM 还是双重犯罪者，以及我们的研究结果是否与 CPORT 第 5、6 和 7 项建议的再犯罪风险水平一致。结果与 HTCU 共享并进行准确性评分。混合模型在 7 个案例中有 6 个是准确的。我们得出的结论是混合模型是可行的，研究结果说明了分析应用程序工件的上下文的重要性。讨论了研究局限性和未来研究建议。