Programme de la conférence

The growing integration of Artificial Intelligence (AI) into European migration governance is reshaping the role and responsibilities of public administrations. This paper critically investigates how AI-based systems are increasingly employed in the surveillance and management of external borders, and explores their implications for fundamental rights, institutional accountability, and the evolving mandate of public authorities across Europe.

Starting from an analysis of the current legal and policy framework, the study maps the adoption of technologies such as biometric recognition systems, risk profiling algorithms, and automated decision-making tools—particularly within EU platforms like ETIAS, EES, SIS, and EUROSUR. These infrastructures allow for the preemptive screening of third-country nationals and the cross-border circulation of personal data, reinforcing a model of digitalized migration management deeply embedded in public administrative practice.

While such systems are often presented as tools for enhancing efficiency and security, the paper highlights the risks they pose to core legal guarantees, including the right to asylum, the principle of non-refoulement, and the right to an effective remedy. The opacity of algorithmic logic, the potential for discriminatory outcomes, and the erosion of procedural safeguards raise crucial questions for European public administrations tasked with implementing and overseeing these technologies.

Drawing on recent jurisprudence from the European Court of Human Rights and the Court of Justice of the European Union, the paper underlines the growing importance of human rights impact assessments and judicial review mechanisms in mitigating the risks associated with high-risk AI systems. Nonetheless, the analysis of the final version of the AI Act (Regulation EU 2024/1689) reveals persistent normative gaps. Despite the regulation’s ambitions, it fails to impose sufficient safeguards on AI applications in migration control, granting broad discretionary powers to public authorities without clear transparency obligations or effective oversight.

The paper argues that this regulatory shortfall reflects deeper tensions within the European public administration landscape—between innovation and rights protection, security logic and legal certainty, efficiency and democratic accountability. It also points to the urgent need for public administrations to develop the institutional capacity, normative awareness, and technical competence necessary to govern AI systems responsibly, especially when they affect vulnerable populations.

By combining legal analysis with public policy reflection, this research contributes to the broader debate on the future of public administration in Europe. It invites policymakers, scholars, and practitioners to critically engage with the digital transformation of migration governance and to shape a regulatory environment that balances technological progress with fundamental rights and the principles of good administration.

This paper aligns with the goals of the EGPA 2025 Conference by offering a forward-looking perspective on the challenges and responsibilities of European public administrations in the age of AI, and by proposing concrete avenues for ensuring legal compliance, ethical use, and institutional resilience in the face of technological change.

This research introduces a novel theory on how selective bureaucratic autonomy arises within prosecutorial agencies in developing, neo-patrimonial, and competitive democracies. We contend that in such settings, political leaders deliberately engineer bureaucratic independence by leveraging incentives—such as increased funding and staffing—to bolster an agency's operational capacity while preserving the outward appearance of autonomy. In reality, this arrangement reflects a tacit agreement: as agencies gain enhanced capabilities and “autonomy,” they concurrently reduce the likelihood of targeting the ruling party or its affiliates.

Our study employs a mixed-methods approach incorporating unprecedented data on municipal-level public prosecutions, electoral backgrounds of politicians, extensive documentary reviews, and numerous interviews conducted in Brazil from 2000 to 2024. Our findings indicate that state governors in Brazil are more likely to grant autonomy and better equip prosecutor offices when targeting opposition municipal politicians. In municipalities led by opposition figures, prosecutorial agencies not only secure increased financial resources but also escalate corruption prosecutions. Conversely, for aligned local politicians, as financial independence and staffing levels rise, fewer corruption cases are investigated.

Contrary to the conventional belief that greater autonomy inherently boosts anti-corruption efforts, our research reveals that political actors can shape autonomy to serve strategic ends while subtly retaining control. This study advances existing literature by challenging the simplistic link between autonomy and anti-corruption effectiveness, and by demonstrating that fiscal independence does not automatically equate to enhanced accountability. Instead, it shows that autonomy and political influence can coexist, as state governors harness institutional resources to direct prosecutorial behaviour. Ultimately, our work contributes new insights into debates on bureaucratic independence and anti-corruption strategies, with significant implications for governance in developing democracies.

Expert governance is a prerequisite for effective governance. Applications for decisions can be quickly assessed for substance and compared with each other and aligned with policy. The experience of repetitive implementation subsequently provides input for recalibrating policy and regulation. Moreover, for various specialised administrative authorities expertise is the main reason for their existence. However, under the liberalisation policies of recent decades, many governmental tasks have been outsourced to the private sector, leaving some expertise absent from the government. For instance, environmental impact assessments almost always turn out to be prepared and assessed by consultancy agencies, because administrative authorities lack that expertise, which leads to major concerns.

At the same time, the law sometimes requires experts involved in decision-making to be independent from the administrative authority. There may be various reasons for this. Sometimes decision-making will have significant consequences for the financial position of the governing body. Possibly the nature of the expertise and decision-making requires that political interference is excluded. It is also possible that the consequences of decision making are so impactful for citizens, and they themselves have so little possibilities to obtain evidence, that requirements of a fair trial demand the expertise to be obtained independently.

This contribution explores the tension between internal and external expert advice and the legal requirements that help rationalise the choice between both. Under what circumstances is the risk of bias so imminent that further requirements should be imposed on expert advice and what do those further requirements entail?

One of the reasons for putting this topic on the agenda is the ECHR's Korošec judgment (ECHR 8 October 2015 (Korošec v. Slovenia, apll.no. 77212/12). In it, the Court ruled that Article 6 ECHR had been violated, because the applicant, claiming a higher disability allowance, was in an unequal procedural position. The crucial medical experts were employed by the administrative authority and the applicant was not able to obtain a contra-expertise himself. The Court ruled that the method of the medical experts’ appointment justifiably gave rise to the applicant’s suspicion that they would not be able to act impartially. That ruling had major impact on Dutch administrative law. The administrative law courts adapted their caselaw, adding new criteria as to whether or not an external expert should be appointed. In most academic discourse, the preference seems to be towards the use of external expertise, as a solid safeguard that an advise is given in complete independence. But is this non-bias always sufficiently guaranteed? External experts are often paid per assignment and may have some financial interest in producing an advice that pleases the client, often the administrative authority.

This paper maps the risk in bias in expertise and the legal requirements that can diminish that risk. Can the quality of expert advice be increased, reducing the risk of bias? Does that work out the same for internal and external experts? That question is addressed using ECtHR and ECJ case law, literature review and the administrative law of the Netherlands as a test case.