The rapid expansion of direct-to-consumer (DTC) genetic testing has generated civilian genomic databases containing tens of millions of profiles, some of which may be accessed, under specific conditions, for forensic purposes. At the same time, artificial intelligence (AI) is increasingly being integrated into forensic genomics workflows, including DNA mixture deconvolution, distant kinship inference, probabilistic phenotyping and investigative lead prioritization. Together, these developments are reshaping the landscape of human identification while raising major scientific, legal and ethical questions.

This presentation provides a multidisciplinary review of the current state of the art at the interface of recreational genetics, forensic genomics and AI. It examines the technological transition from conventional STR-based forensic profiling to SNP-based approaches compatible with investigative genetic genealogy, as well as the growing contribution of next-generation sequencing and machine learning tools in complex forensic interpretation. Particular attention is paid to the analytical strengths of AI-assisted approaches, but also to their limitations, including data quality dependence, population bias, lack of standardization, uncertainty in long-range kinship inference and the explainability requirements associated with high-risk AI systems.

The presentation also addresses the fragmented legal and regulatory environment governing the forensic use of recreational genetic databases, with a focus on France, Europe and international practices. Key bioethical issues include consent, proportionality, privacy of genetic relatives, algorithmic discrimination and democratic oversight of genomic data use in criminal investigations.

The central message is that the forensic use of recreational genetic databases cannot be assessed solely as a technical innovation. Its long-term legitimacy and operational viability depend on a structured alliance between robust genomic science, validated AI tools, clear legal governance and strong ethical safeguards. This emerging field has considerable potential for human identification, but only if scientific performance is matched by transparency, proportionality and accountability.

Authors

• Christian SIATKA (UPR-CHROME, Faculté des Sciences, Place Gabriel Péri, France)
• Francis HERMITTE (Institut de Recherche Criminelle de la Gendarmerie Nationale, IRCGN, France)
• Sylvain HUBAC (Institut de Recherche Criminelle de la Gendarmerie Nationale, IRCGN, France)
• Stéphane SAUVAGERE (École de l'ADN, France)