Protect Not Surveil

Protect Not Surveil is the framing and hashtag #ProtectNotSurveil through which a 300+-organisation coalition — anchored by Access Now, EDRi (European Digital Rights), the Platform for International Cooperation on Undocumented Migrants (PICUM), and the Refugee Law Lab — advocates for the prohibition of specific artificial intelligence applications at European borders and in EU migration enforcement. The framing asserts that people exercising their right to move, seek safety, and pursue opportunity are entitled to do so without algorithmic profiling, predictive risk-scoring, biometric identification, and the forms of discrimination these systems deliver; and that the AI applications it names cannot be made safe through technical fixes or procedural safeguards and must therefore be banned. It launched in April 2023 as the EU AI Act entered its final legislative stages and has continued since, most recently as opposition to the proposed new Europol Regulation, which the coalition argues extends the same rights-violating surveillance logic the AI Act failed to prohibit.

Origin

The framing formed in February 2023 and publicly launched in April 2023 — the moment when the EU AI Act's trilogue negotiations between the European Parliament, the Council, and the Commission were reaching their decisive final phase. The coalition's entry into the policy contest was late relative to the broader EDRi-led AI Act civil-society coalition, which had operated since the Commission's April 2021 proposal; the #ProtectNotSurveil launch narrowed the framing specifically to migration and border contexts, giving that dimension of the AI Act debate its own named coalition, hashtag, and advocacy surface.

The campaign's founding four organisations each brought distinct institutional standing to the coalition. Access Now brought its international digital-rights network and established record in EU policy campaigns. EDRi brought its coordinating role across European civil society and its infrastructure from the parallel Reclaim Your Face campaign against biometric mass surveillance in public spaces — a campaign that had run since 2020 and with which #ProtectNotSurveil shared overlapping demands and several founding organisational members. PICUM brought direct expertise and relationships with the migrant-rights and refugee-advocacy communities whose members were the population most directly subject to the AI systems the campaign opposed. The Refugee Law Lab brought legal expertise in refugee and international protection law, giving the coalition's demands their grounding in binding international legal obligations rather than solely in the EU's Charter of Fundamental Rights.

The core argument

The framing makes four categorical demands. First, it calls for the prohibition of predictive analytics systems used to prevent migration: systems that forecast migration flows or predict individual likelihood of irregular crossing and use those predictions to deploy border-enforcement resources or trigger administrative action. The coalition's argument is that these systems systematically over-represent racialised and disadvantaged populations in high-risk output because they are trained on historically biased enforcement data, and that any algorithmic system designed to prevent people from exercising their right to seek asylum is unlawful as a matter of international refugee law regardless of its technical implementation.

Second, it calls for a prohibition on automated risk-assessment and profiling systems in migration and border contexts — systems that assign risk scores to individuals crossing borders, flag individuals for additional scrutiny, or generate profiles used by border officials to make admissibility or processing decisions. The campaign's argument here mirrors msg-ban-biometric-mass-surveillance's reasoning about biometric surveillance in public spaces: that the systems' outputs embed racial and geographic proxy discrimination that procedural safeguards cannot remove, and that the use of algorithmic outputs in consequential migration decisions violates the right to individual assessment that EU and international law require.

Third, it calls for the prohibition of pseudo-scientific "lie-detector" and emotion-inference technologies deployed at borders — AI systems claimed to detect deception, emotional state, or behavioural anomaly from physiological or facial signals, which several EU member states piloted at border crossing points. The coalition characterises these as scientifically baseless and regards their deployment as a per se rights violation independent of any procedural wrap.

Fourth, it calls for a prohibition on remote biometric identification at border crossing points and in and around immigration detention facilities — a demand that overlaps directly with the Reclaim Your Face campaign's broader demand for prohibition of remote biometric identification in publicly accessible spaces, but which the #ProtectNotSurveil campaign grounds specifically in the vulnerability of detained and mobile populations and in the absence of meaningful consent in border-crossing contexts.

The campaign's framing is explicitly abolitionist on these four points: it argues, at the April 2023 launch, that these AI systems are "racist and lethal," that they "will never be fixed either via technical means nor with some procedural safeguards," and that regulation rather than prohibition is insufficient. This places the framing to the left of much digital-rights advocacy, which often accepts regulation-plus-safeguards as achievable within the EU legislative framework.

Key moments

The EU AI Act's March 2024 plenary adoption was the coalition's first major verdict event. The coalition's joint post-adoption statement, signed by Privacy International, EDRi, Access Now, and coalition partners, characterised the adopted text as "a dangerous precedent" for migrant rights: while the Act included partial restrictions on remote biometric identification and partial bans on emotion recognition in workplaces and educational settings, it did not include the prohibitions on migration-context profiling and predictive analytics the campaign had sought. The migration carve-outs — which allowed law-enforcement and border-management authorities wider latitude than other public-sector users — were identified as the Act's central failure from the campaign's perspective.

The April 2024 EU Migration and Asylum Pact brought a second mobilisation: the coalition warned that the Pact, which mandated expanded digital data collection and interoperability across EU border and migration databases, was "ushering in a deadly new era of digital surveillance" directed at people on the move, and that the Pact's database-expansion provisions would amplify the harms the AI Act had failed to prevent.

The proposed new Europol Regulation brought the campaign's most recent sustained opposition. The February 2025 #ProtectNotSurveil position paper on Europol — signed by 120+ organisations — argued that the proposed regulation would grant Europol expanded authority to process biometric data and run predictive analytics directed at migration, extending the surveillance infrastructure the AI Act had left partially in place. In November 2025, the European Parliament's LIBE Committee voted to advance the regulation despite the coalition's opposition; EDRi characterised the vote as "a dangerous step towards mass surveillance in the EU" and announced continuing mobilisation ahead of the Parliament's full plenary vote.

Why it has carried

The framing's durability across the EU AI Act, the Migration Pact, and the Europol Regulation rests on three features.

First, it names the population the European digital border apparatus is directed at — migrants, refugees, and people in precarious movement — rather than the general public who are the subject of most digital-rights campaigning. This specificity makes the rights-violation argument concrete: the campaign does not claim that AI at borders harms everyone indiscriminately, but that it harms a specific, legally protected population in specific, disproportionate ways that international and EU law prohibit.

Second, it maintains a stable abolitionist demand across shifting legislative vehicles. Where many civil-society campaigns adapt their demands to each legislative text, accepting partial wins and recalibrating, #ProtectNotSurveil has held the prohibition demand constant across the AI Act, the Migration Pact, and the Europol Regulation — which means it can function as the oppositional framing for any EU legislative instrument that extends migration-context AI surveillance, without requiring each campaign to build a new analytical frame from the beginning.

Third, the coalition's multi-sector composition — digital rights, migrant rights, refugee law, algorithmic accountability — makes the framing legible across advocacy communities that do not normally operate from a shared frame. A digital-rights organisation and a refugee-law clinic share the #ProtectNotSurveil hashtag without needing to adopt each other's full analysis; the shared framing creates coalition surface that a narrower frame would not.

The framing's direct structural parallel is msg-no-tech-for-ice — the 2018 Mijente-launched US framing against tech-company contracts with ICE. The two operate in different jurisdictions and with different target audiences (EU institutions and member states for #ProtectNotSurveil; private corporations and the US tech-talent pipeline for #NoTechForICE), but share the abolitionist orientation toward specific surveillance uses in migration enforcement and the argument that the corporate and institutional supply of migration-surveillance technology makes suppliers responsible for its harms.