VANGUARD adopts a comprehensive approach to Trafficking in Human Beings (THB), encompassing both its offline and online manifestations and the links between them. The growing use of digital technologies has created new opportunities for perpetrators across different stages of the trafficking process. In response to these evolving threats, VANGUARD seeks to leverage technological developments to support the prevention and detection of THB, assist police and border guard authorities in identifying and prosecuting perpetrators, strengthen support for victims, and contribute to information dissemination and awareness-raising.
The technologies and research activities developed within VANGUARD are conceived and assessed against the applicable legal framework from the earliest stages of development. This framework comprises the Charter of Fundamental Rights of the European Union, the AI Act, the General Data Protection Regulation (GDPR) and the Law Enforcement Directive (LED). Fundamental rights considered particularly vulnerable to interference in the law enforcement context include human dignity, equality and non-discrimination, privacy and data protection, freedom of expression and information, the right to an effective remedy and a fair trial, and the presumption of innocence. Risks to these rights are identified by reference to the functionalities and intended uses of each tool and are reviewed through regular fundamental rights impact assessments. Although such assessments are not mandatory during the development phase, VANGUARD conducts them regularly to assess compliance with applicable legal and ethical requirements and to anticipate the legal and operational needs of end users.
Legal assessment accompanies technical development and anticipates deployment needs, rather than being conducted only after a tool is completed. The assessment covers: (i) choices relating to data, data processing, and data governance; (ii) technical documentation and quality management; (iii) record-keeping and logging; (iv) transparency and information to deployers; and (v) corrective actions. It further examines how outputs are generated and communicated, including the use of confidence information and appropriate limitations to avoid presenting probabilistic results as definitive findings, as well as aspects that may need to be addressed in the instructions for use. Finally, it takes into account the legal and operational obligations of police and border guard authorities deploying the tools.
VANGUARD develops both high-risk and non-high-risk AI systems to detect and identify online and offline-facilitated THB. Both categories of systems are assessed against the ethical requirements identified in the Ethics Guidelines for Trustworthy AI: (i) human agency and oversight; (ii) technical robustness and safety; (iii) privacy and data governance; (iv) transparency; (v) diversity, non-discrimination and fairness; (vi) environmental and societal well-being; and (vii) accountability. Although these ethical requirements are not legally binding, they closely align with several mandatory requirements applicable to high-risk systems under the AI Act. The cumulative application of legal and ethical requirements, which pursue broadly aligned protective objectives, therefore often results in comparable safeguards for both high-risk and non-high-risk systems, including where the corresponding AI Act requirements do not legally apply.