AI-driven efficiency at what cost? Rethinking EU migration policy in the digital age

Has the EU become overly reliant on technological solutions to what are fundamentally political and humanitarian challenges? Not necessarily – but only if it chooses to reform how it funds them.

The European Union has become a leading actor in utilising digital technologies to strengthen border control and mitigate security risks. Used well, artificial intelligence (AI) can help standardise asylum decisions and ease pressure on overburdened border officials, particularly those in southern European countries experiencing a major influx of refugees in recent years.  The technology itself is not the problem; the issue is how the EU chooses to fund and govern it.

The European Union’s Multiannual Financial Framework (MFF) for 2021-2027 has established a budgetary cap of €22.7bn for migration and border management.^[1] This financial allocation indicates a significant pivot towards digital solutions as part of border management policy; in fact, AI migration technologies are already being employed by national governments in 11 member states. Further marking this shift is the Integrated Border Management Fund (IBMF), which includes the Border Management and Visa Instrument (BMVI) and the Customs Control Equipment Instrument (CCEI). The BMVI alone has received a dedicated budget of €6.2bn, a 135% increase from the previous budget.^[2] While these are clear strides in Europe’s digital sector, and a step closer to Europe’s goal of “strategic autonomy”, a closer analysis of these allocations reveals a troubling asymmetry. Enormous resources are dedicated to surveillance infrastructure, while comparatively little goes toward humanitarian assistance, drawing significant criticism from groups concerned with the fundamental rights of EU citizens and migrants alike.

Shifting EU funding priorities

Just over a third of the migration and border management budget is earmarked for the development and expansion of databases and information systems. This is crucial for many new programmes, such as the European Travel Information and Authorisation System (ETIAS), set to launch later this year. ETIAS will serve as a pre-travel screening system that determines eligibility for short stays in the EU, thereby adding another layer to the EU’s border security system.^[3] By contrast, only 0.04% of the same funding is allocated to projects supporting migrants in need of international protection.

Each MFF cycle is heavily contested and requires approval from all 27 member states. So why is there such a pronounced emphasis on technology now? The answer lies partly in the sheer scale of international mobility and global migration flows. Proponents of AI argue that it can significantly enhance the efficiency and accuracy of  migration management, from processing visa and residence permit applications to handling travel authorisations and asylum claims.^[4] They also contend that the use of AI can improve neutrality and standardisation in decision-making, thereby reducing the risk of discriminatory treatment.

 The limits of digital border governance

In principle, these benefits are deeply impactful. However, critics warn that key concerns and considerations are being overlooked when these digitalised border policies are put into practice. For instance, the EU Artificial Intelligence Act (AI Act), the world’s first-ever comprehensive legal framework on AI, seeks to provide regulatory oversight for so-called ‘high-risk’ AI systems, a category that includes many border management tools. Yet civil society critics, including EuroMed Rights, Platform for International Cooperation on Undocumented Migrants (PICUM) and European Digital Rights (EDRi), describe the legislation as weak. They argue that algorithmic bias remains a serious risk: biometric identification systems, for instance, have demonstrated lower accuracy rates for certain ethnic groups, undermining the very neutrality they are meant to provide.^[5] Acknowledging this risk, the AI Act itself classifies AI systems used to scan areas of the border and deter migrants as high-risk, and prohibits real-time remote biometric identification for law enforcement purposes in publicly accessible spaces.^[6]

These debates are happening in real time, even as border management policies using AI are deployed. As of 12 June, all EU member states must have implemented the Pact on Migration and Asylum. The Pact attempts to establish a common EU-wide system and standards when it comes to managing migration. Although celebrated by EU policymakers after two years of negotiations, it has been broadly criticised by civil society and specifically the #ProtectNotSurveil Coalition[7], which claims the Pact is a “dangerous regime” of migrant surveillance.^[8]

What’s more, the EU recently approved a new ‘Returns Regulation’ this month – commonly referred to as the ‘Deportation Regulation’ by civil society – inviting fresh criticism. The Regulation promises faster returns, dramatically expands the ‘use and duration of immigrant detention’[9], normalises ‘ICE-style’ police raids and relegates the protection of human rights to an “afterthought”[10]. Importantly, critics have flagged that this Regulation is part of the EU’s wider ‘digital deportation machine’[11], which relies on the collection and sharing of migrants’ personal data, thereby extending surveillance beyond Europe’s borders.

 Gaps in the current policy framework

The prioritisation of security over humanitarian support raises serious ethical concerns, yet few serious analysts argue that technology should be removed from migration governance entirely. AI integration into multiple aspects of our world is inevitable, and the necessity to manage the movement of millions of people means migration governance is not excluded from this.^[12]

While AI-driven border technologies may improve administrative efficiency, current EU migration policy places disproportionate emphasis on border management systems while providing comparatively limited investment in the social and legal support structures that migrants rely upon. This imbalance undermines the EU’s stated commitments to democracy and fundamental rights.

As negotiations on the 2028–2034 MFF are underway, the European Commission should establish a mandatory minimum allocation dedicated to legal aid and integration support. When the current framework allocated 36.30% to surveillance infrastructure and just 0.04% to migrant protection, a minimum of 5-10% is not an ambitious target. This would not require reducing investment in border technology; it would simply ensure that significant spending on surveillance is matched by a proportionate commitment to the people interacting with these systems. Furthermore, the AI Act establishes that high-risk AI systems require rigorous oversight, but there is a lack of transparency in the progress of implementation. Rather than relying primarily on pre-deployment conformity assessments, the EU should require ongoing accountability for AI systems used in border management. Annual fundamental rights impact assessments should be a condition of continued EU funding, helping to translate the AI Act’s principles into effective oversight while ensuring that innovation can continue within a robust framework for protecting fundamental rights.

The European Parliament and Council of the EU are set to negotiate the European Commission’s MFF proposal ahead of the 2028 deadline, making the next 18 months the critical window for these reforms to be embedded in the final budget.

The EU is not yet over-reliant on technologies. However, without a clearer balance between investment in border technologies and investment in legal protection and integration, it risks building a migration system that is highly efficient in administrative terms but weak in safeguarding rights. The next MFF offers a rare opportunity to get things right.

^[1] European Commission, EU Multiannual Financial Framework 2021–2027, EUR-Lex, https://eur-lex.europa.eu/EN/legal-content/summary/eu-multiannual-financial-framework-2021-2027.html.

^[2] PICUM, Beyond Walls and Fences: EU Funding Used for a Complex and Digitalised Border Surveillance System, July 2024,https://picum.org/wp-content/uploads/2024/07/Beyond-walls-and-fences_EU-funding-used-for-a-complex-and-digitalised-border-surveillance-system_EN.pdf.

^[3] ETIAS, Who Needs ETIAS?,https://etias.com/what-is-etias/who-needs-etias.

^[4] Verfassungsblog, Regulating AI at Europe’s Borders,https://verfassungsblog.de/regulating-ai-at-europes-borders/.

^[5] Ibid.

^[6] European Commission, Regulatory Framework for AI,https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai.

[7] #ProtectNotSurveil, https://protectnotsurveil.eu/.

^[8] European Commission, Regulatory Framework for AI,https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai, p.2.

[9] PICUM, New Returns Regulation ushers in dystopian detention and deportation regime, https://picum.org/blog/new-returns-regulation-ushers-in-dystopian-detention-and-deportation-regime/.

[10] EESC, Return regulation: ice-cold prospects for human rights in Europe, https://www.eesc.europa.eu/en/news-media/press-releases/return-regulation-ice-cold-prospects-human-rights-europe.

[11] Rodelli, Catherine, Stefi Richani & Hope Barker, How the EU is building a dystopian, surveillance-driven, deportation machine, https://euobserver.com/205537/how-the-eu-is-building-a-dystopian-surveillance-driven-deportation-machine/.

^[12] IOM, World Migration Report 2026: International Migrants — Numbers and Trends, Chapter 2,https://worldmigrationreport.iom.int/what-we-do/world-migration-report-2026/chapter-2/international-migrants-numbers-and-trends.

The views expressed in this #CriticalThinking article reflect those of the author(s) and not of Friends of Europe.