Hamoudi v Frontex, an EU Courts pushback case: Shifting the burden of proof and a duty to assist the Court (a duty of candour?)

 

Antje Kunst*

* Antje Kunst is an international lawyer and barrister of Garden Court North Chambers, admitted to the Bar of England and Wales, and the Bar of Berlin, advising and representing individuals in a wide range of matters related to fundamental rights within the CFSP and other fields. She has appeared in numerous cases before both the Court of Justice and the General Court, within the Court of Justice of the European Union.  

Photo credit: Rock Cohen, via Wikimedia Commons

 

Introduction

On 4 February 2025 the Grand Chamber held a hearing on the appeal of Hamoudi v Frontex in Case C-136/24 against the General Court (GC)’s Order of 13 December 2023. This appeal is taking place alongside the separate challenge in WS v Frontex (see analysis of that case and summary of the hearing).  

The case of Hamoudi v. Frontex concerns Syrian asylum seeker Alaa Hamoudi, who alleges that on 28 April 2020, he and 21 other individuals were subjected to a pushback operation in the Aegean Sea. He claims that upon arriving from Turkey by boat, he entered Greek territory on the island of Samos to seek asylum. Local police intercepted him, confiscated his mobile phone, and later that day, Greek authorities forced him and the others back out to sea. The following day, a vessel from the Turkish coast guard took them aboard. Hamoudi also alleges that on 29 April 2020, while at sea, a private surveillance aircraft operated by Frontex, equipped with a camera, flew over the scene twice.

In February 2022 OLAF issued a highly damaging report on Frontex which addresses incidents of illegal pushbacks involving Frontex assets like Frontex Surveillance Aircraft, in particular in the Aegean Sea, in late April 2020. This report was not made publicly available (see here more on this fact).

In March 2022, Mr. Hamoudi brought an action pursuant to Article 340.2 TFEU  against Frontex, seeking €500,000 compensation for non-material in damage in respect of violations of his fundamental rights, including the prohibition of collective expulsion and the principle of non-refoulement under the EU Charter of Fundamental Rights.

The GC dismissed the action by Court Order stating it "manifestly lacked any foundation in law." (para. 62 of the Court Order). The GC found that the appellant’s own written statement taken by an NGO more than a year and a half after the events, a Bellingcat article, and four screenshots from third-party video recordings of the pushback were “manifestly insufficient” to conclusively prove Hamoudi’s presence or involvement in the incident.  It further noted that the claimant could not be identified in the images. Further his own written statement lacked credibility (paras. 40 and 41 of the Court Order).

Hamoudi appealed this decision on 19 February 2024, contending that the General Court erred in its legal assessment and mischaracterized the facts.

Relevance of two ECtHR expulsion cases

In January 2025, in two groundbreaking judgments A.R.E. and G.R.J. v. Greece the European Court of Human Rights (ECtHR) unanimously confirmed Greece's “systematic practice” of pushbacks of third-country nationals from Greece back to Turkey. Both ECtHR cases are very similar in terms of facts to the Hamoudi case.  The G.R.J.  case which was referred to by the judges during the hearing concerns like in the Hamoudi case a pushback from the island of Samos to Turkey in 2020.

Significantly the parties were asked to comment on the relevance of these two judgments seemingly with a view to consider the ECtHR’s finding on the existence of a “systematic practice” of pushbacks by Greece. The Court might likely to be inspired by the ECtHR approach in these types of cases, if not align its case law with the Strasbourg Court’s large case law on expulsion. Both Courts have engaged in the past in a judicial dialogue, referring to each other’s case law, aiming for consistency in protecting human rights of individuals.

Assisting the court in reaching the correct result

Pushbacks are difficult to prove, in particular when they happen at night and mobile phones are confiscated or destroyed (as happened to the appellant and applicants in ECtHR cases (e.g., in  A.R.E.  para. 266).

In proceedings before the ECtHR, respondent states often deny the facts rather than disclosing relevant records, surveillance footage, or photos and videos documenting the events, as elaborated here. In Hamoudi, Frontex representatives, during the hearing before the Grand Chamber, claimed it was unclear whether Frontex had been present at the scene of the events, noting that "they had not been monitoring the situation from afar through their binoculars" (see here).

The question arises: Is it acceptable for an EU agency like Frontex to present such arguments, displaying a stance similar to that of states before the ECtHR? Or does it, instead, have a duty to assist the Court? Article 24 of the Court’s Statute provides that the Court “may require the parties to produce all documents and to supply all information which the Court considers desirable.” This provision establishes not only a procedural obligation for the parties to cooperate fully with the Court but also an implicit duty to assist the Court in reaching an accurate outcome. The Court here relies on the transparency and good faith of the parties to uncover the relevant facts and produce it before the Court.

In UK public law there is a duty of candour vis-à-vis the Court. The duty of candour requires a public authority "not to seek to win [a] litigation at all costs but to assist the court in reaching the correct result and thereby to improve standards in public administration." (see also here)

This principle is not unknown in member states. In Germany for example pursuant to §99 Administrative Court Procedure Code and §138 Code of Civil Procedure , public authorities must act truthfully and fully cooperate with the administrative courts.

Arguably as an EU agency, Frontex, and based on Article 24 of the Court’s Statute it has a duty to assist the Court in reaching the correct result by providing information and disclosing evidence exclusively within its possession, even if that evidence may be unfavourable to its case. In other words, it is not acceptable for Frontex to adopt the same approach as states in these types of cases.

Furthermore, contrary to what the EU Agency presented before the Court, Frontex also “sits on evidence” -not only the member states- related to its activities collecting evidence when it carries out its assistance and surveillance activities acting under a clear mandate to protect the fundamental rights of individuals in distress at sea, in accordance with Articles 80(2) and 80(3) of the Frontex Regulation.

Impossible proof

The burden of proof imposed by the GC amounted to probatio diabolica, an impossible proof for Mr. Hamoudi.

There might be cases in which applicants are able to provide robust evidence related to their individual situation, such as photographs, video recordings, and witness testimonies. The possession of mobile phones with geolocation capabilities gives victims of pushback operations new means to provide proof of their presence in a certain area or to prove the involvement of Frontex.

In most cases, individuals like Hamoudi face serious practical difficulties in gathering evidence. The incident took place at night and under very stressful conditions, with Hamoudi being in the open sea in a boat. Hamoudi was unable, both physically and mentally, to gather any evidence. (see more on this here) Taking away mobile phones as happened to Mr. Hamoudi makes it impossible for the victim to use such evidence.  This is what the General Court failed to consider.

Applying the rules governing the burden of proof in Strasbourg expulsion cases

The appellant argues in his appeal that the General Court (GC) committed an error of law by failing to properly assess and apply the relevant rules governing the burden of proof. Specifically, the appellant contends that the GC did not consider the legal criteria established by the ECtHR in its expulsion case law, which aligns with the EU Court’s own established practice (e.g., in discrimination cases).

When adjudicating such cases, the ECtHR shifts the burden of proof to the state as elaborated here  where there is different access to information which advantages the state and leaves the applicant without evidence or when the defending State possesses information that could  corroborate or refute the applicant’s allegations. Despite what Frontex alleged at the hearing Frontex certainly has access to information regarding the events in question as also the OLAF report shows. This relates to information regarding its own compliance with its own fundamental rights obligations and the existence of the immaterial harm stemming from the events during the night in question.

The ECtHR expulsion case law provides that once an applicant has furnished prima facie evidence in support of his or her version of events, the burden of proof should shift to the respondent when (a) there is an absence of personalized treatment—such as not being interviewed or having personal details taken—which lies at the very core of the applicant’s complaint, as in the present case; and (b) this absence has contributed to the difficulty in adducing evidence of involvement in the event (see Case of N.D. and N.T. v. Spain, para. 85).

As analysed here in its recent rulings the ECtHR has explicitly acknowledged that the state’s complete denial of alleged facts places the applicant in an inherently difficult evidentiary position, in which they may be unable to establish the veracity of their account (A.R.E., para. 218; G.R.J., para. 183).

The consolidated case law of the ECtHR regarding the reversal of the burden of proof might serve as a compelling source of inspiration for the ECJ’s judges in this specific case, which bears significant similarities to the large number of expulsion cases the ECtHR has adjudicated, most recently in G.R.J.

The fact that this is an application for damages under Article 340(2) TFEU against an EU agency, rather than a human rights application before the ECtHR against a state, does not make a difference, as similar objectives are pursued. The judges might consider that the conditions to be met for being successful with claim for damages before the ECJ are significantly higher than a human rights application before the ECtHR. This applies particularly to the more extensive requirements for proving harm under the ECJ’s case law compared to the standards for establishing victim status under Strasbourg case law. In other words, and contrary to what Frontex appeared to imply during the hearing, the bar for establishing liability remains high and substantial.

Meet prima facie threshold

To meet the prima facie threshold under the ECtHR, regularly two key elements are required (a) an individual account which is specific, consistent, generally coherent, and credible and (b) general context evidence concerning the broader context relevant to applicants’ claims.  In G.R.J.  the ECtHR relied on such general context evidence: “[h]aving regard to the large number, diversity, and concordance of the relevant sources …the Court concludes that there is serious evidence to suggest that, at the time of the alleged events, there was a systematic practice of refoulement by the Greek authorities of third-country nationals from the Greek islands to Turkey.” G.R.J. , para. 190, unofficial translation. Contradictions in the respondent’s statements will be considered as well.

It appeared from the judges' questions during the hearing that the Court was indeed considering holding that the General Court had failed to consider shifting the burden of proof to the EU Agency. This is not surprising in view of its own well-established case law on the reversal of the burden of proof in similar cases of asymmetry of access to information, albeit in other areas (e.g., discrimination) and its approach to consider relevant case law of the ECtHR as a source of inspiration. In this regard, the Court has ample reasons to hold that the applicant has provided prima facie evidence of the violation and that the General Court should have considered shifting the burden of proof to Frontex.

The judges inquired about OLAF’s 2021 report on Frontex and its findings concerning the events of 28 and 29 April 2020. It mentioned the fact that the report refers to two Frontex officers confirming the credibility of the Bellingcat article admitting that the operation of which Mr. Hamoudi claims to be the victim of took place. In any event, if the case had been considered by the Strasbourg Court, it would have found that the specific and consistent account by Mr. Hamoudi, partially corroborated by OLAF’s findings and the Bellingcat article, constitutes prima facie evidence, thereby triggering a shift in the burden of proof.  The Court of Justice would be justified in following the Strasbourg case law and reaching this conclusion. This is also so because of the widely documented information on the general context and the myriad of contradictory statements about the events by Frontex outside the proceedings.

Refuting the appellant’s claims

Frontex stressed at the hearing that it could not “live up” to substantiating, refuting or corroborating the prima facie evidence of the appellant. It was in a situation of an impossible proof. This is wholly unconvincing. Frontex could simply review its records or surveillance footage to determine whether, at the relevant date and time, its surveillance aircrafts were operating over the area where the unlawful pushback of Mr. Hamoudi occurred. Frontex conducted two active operations in the region and is best placed to substantiate or refute the appellant’s claims. With an explicit to mandate to provide support to Greece in the Aegean Sea fully aware of Greece’s shady practices and under an explicit mandate to safeguard fundamental rights of individuals in distress at sea as per Article 80 (2) and 80 (3) of the Frontex Regulation it was monitoring the situation.

Conclusion

Enabling the EU Courts to focus on the legal assessment of alleged fundamental rights violations in such cases by providing as much factual information as possible serves the interests of all parties involved—including the Court itself.

Frontex has faced significant criticism over alleged pushback practices in the Aegean Sea, raising serious concerns about compliance with its fundamental rights obligations. It is in the Agency’s interest to be as transparent as possible, not to seek to win a litigation such as the Hamoudi case at all costs but to assist the court in reaching the correct result. By doing so, Frontex can contribute to reinforcing trust in the EU Agency’s commitment to human rights and accountability.

WS and Others v Frontex before the Grand Chamber: Ensuring Meaningful Protection of Fundamental Rights in Forced Returns by an EU Agency acting as a Safety-Net

 

Antje Kunst*

*Antje Kunst is an international lawyer and barrister of Garden Court North Chambers, admitted to the Bar of England and Wales, and the Bar of Berlin, advising and representing individuals in a wide range of matters related to fundamental rights within the CFSP and other fields. She has appeared in numerous cases before both the Court of Justice and the General Court, within the Court of Justice of the European Union.  

Photo credit:
Влада на Република Северна Македонија via Wikimedia Commons

 

Introduction

On 4 February 2025 the Grand Chamber held a hearing on the appeal in Case C-679/23 P against the General Court (GC)’s Judgment of 6 September 2023 in case WS and Others v. Frontex, Case T- 600/21. The case's high significance lies in the fact that the Court of Justice looks the first time at Frontex's liability in the context for illegal return operations. This article provides a brief summary of the case and comments on some of the issues discussed at the hearing. (For a summary of the arguments at the hearing, see here).

Despite high expectations by many regarding the Grand Chamber’s ruling in this case, the Court may only be asked to rule pursuant to Article 256 TFEU and Article 58 of the Statute of the Court of Justice on whether the GC committed an error of law. Thus, the judicial review might be limited in this respect.

Given the issues discussed at the hearing, it is expected that the Court of Justice will clarify Frontex’s obligations in joint return operations based on the circumstances of the individual case on appeal. Whether it will extend its reasoning beyond this case remains to be seen. It would certainly be desirable if the Court elaborated on Frontex's own fundamental rights obligations, which are distinct from Member States' obligations, also to counter the tendency of the Agency to negate its own obligations, as was noticeable at the hearing. One can certainly expect that Advocate General Ćapeta will offer interesting insights in her Opinion, which she is set to deliver in June 2025.

As a factual background the applicants, a Syrian Kurdish refugee family with four children, arrived on the Greek island of Milos in October 2016 by boat with the intention of claiming asylum. Even though they made their intention clear to apply for asylum just six days later, they were deported by flight to Turkey in a joint return operation conducted by Greece and Frontex. Because of the fear of being returned to Syria by the Turkish authorities (para. 59 of the GC’s judgment), the family decided not to stay in Turkey but to proceed to Iraq, where they reside to this day.

The applicants filed a claim for compensation, alleging that Frontex had violated its obligations under the then-applicable Frontex Regulation (the 2016 Frontex Regulation) and several fundamental rights during the return operation. Specifically, they claimed Frontex failed to comply with its obligations related, inter alia, to the principle of non-refoulement, the right to asylum and the prohibition of collective expulsion.

Admissibility of the action: Article 340.2 TFEU

Any arguable claim that fundamental rights have been violated should be subject to an effective judicial remedy, and that is in these types of cases a claim for compensation under Article 340.2 TFEU (see  more here). Despite Frontex’s arguments, the GC rightly deemed the claim admissible, which some view as a significant success in itself, as it sets a precedent for the admissibility of similar actions.

The GC’s (implicit) conclusion regarding Frontex’s verification duties

The GC dismissed the action finding that Frontex's role in joint return operations was to provide technical and operational support to EU Member States and not to enter the merits of return decisions or decide on applications for asylum (paras. 64 and 65 of the GC’s judgment). 

The GC’s holding that Frontex cannot assess the merits of return decisions or decide on asylum applications is beside the point. By failing to distinguish between a return decision and its implementation, the GC ignored Frontex’s role in the implementation of return decisions. The implementation of returns clearly falls within Frontex’s competence, and this was precisely what the appellants were challenging before the GC. That is why the GC’s findings in this regard constitute an error in law as pleaded by the appellants (see the appeal).

An EU Agency implementing member states’ return decisions

The 2016 Frontex Regulation requires the EU agency to provide necessary assistance and coordinate or organize joint return operations together with member states. With this role come specific responsibilities, which arise undoubtedly not only after a joint return operation, but also before and during such an operation, especially in monitoring and ensuring that individuals like the Syrian refugee family in question are protected from fundamental rights violations.

The 2016 Frontex Regulation specifically requires the agency to ensure respect for fundamental rights, including the principle of non-refoulement, throughout the operation. Its monitoring role in relation to return operations is outlined in Article 28(7) of the 2016 Frontex Regulation, which obligates Frontex to report any fundamental rights concerns to the participating Member States. Additionally, Article 34(2) states that Frontex must ensure no person is returned to a country where they could face refoulement. As an EU agency, it must follow the EU Charter of Fundamental Rights. The EU Return Directive further sets out rules on enforcing return decisions, including postponing a return if there is a risk of refoulement.

Independent verification duties to ensure protection of fundamental rights

From the questions asked it appears that the Court will take the view – quite correctly - that Frontex is, at a minimum, under an obligation to check if it is assisting, organising or coordinating the return of actual returnees, i.e. persons who are subject to written return decisions. It cannot simply trust the national authorities and turn a blind eye to who is on the returnees list.

In its own submissions at the hearing, Frontex acknowledged a duty to verify cases where the national documentation, including written return decisions, contains a logical or a manifest error.

The absence of enforceable return decisions in this case is not disputed by Frontex. In this respect, Frontex failed to check the relevant national paperwork. A Syrian-Kurdish family was returned where they came from, Turkey, despite having expressed their intention to seek asylum in Greece - a fact that a proper verification of the national paperwork would have likely revealed.

Frontex should have identified a clear error on the returnees list, as there were sufficient indications that the Syrian-Kurdish family, including four children, had been wrongly included. The presence of this family on the return flight—after arriving in Greece by boat and being scheduled for return a few days thereafter to Southeast Turkey without apparently having filed an asylum claim—should have raised concerns or suspicion about their possible wrongful inclusion on the list of returnees.

Regarding the risk of refoulement, the GC acknowledged that the appellants feared "being returned to Syria by the Turkish authorities" (para. 68 of the GC’s judgment). In this context, the GC failed to rule that Frontex is obligated to assess whether individuals like the Syrian-Kurdish family risk being transferred to a third country where they may face refoulement or violations of their fundamental rights. Frontex has a duty to prevent any form of refoulement in line with its own fundamental rights obligations as clearly set out in the Frontex Regulation. 

The GC failed to rule that Frontex could and should have enquired, for example by contacting the Greek authorities why this family was on the return flight to Turkey and whether their placement on the list was justified. The Agency’s failure to do so constitutes an unlawful act which the GC failed to establish.

Frontex submitted at the hearing that a few years after the events in question, a new IT tool was introduced by Frontex, requiring Member States to confirm several key conditions before a return operation can proceed. These conditions include: (a) the existence of an individual implementable final return decision, (b) the exhaustion of all judicial remedies, (c) confirmation that the returnee has not applied for asylum. To fulfil in a meaningful way its fundamental rights obligations, Frontex must actively verify this information or, at the very least, implement robust mechanisms to ensure that these criteria are met before facilitating or participating in return operations. It failed to do so in the present case.

Consequently, the GC’s conclusion that Frontex had no obligations regarding the return flight of the family to verify, to the very least, whether the return operation involved actual returnees (see the appeal) constitutes an error of law.

Sufficiently direct causal link

The GC concluded that Frontex could not be held liable for any damages related to the return operation, as the appellants had not adduced evidence of a sufficiently direct causal link between the damage invoked and the alleged conduct of Frontex (para.71 of the GC’s judgment).

The appellants correctly contended that if Frontex had fulfilled its obligations regarding the protection of fundamental rights in the context of joint operations, they would not have been unlawfully returned to Turkey. Instead, they would have obtained the international protection to which they were entitled, given their Syrian nationality and the situation in Syria at the time.

In the appeal, the appellants argued that the GC erred in law by concluding that the necessary causal link between Frontex’s actions and omissions and the harm suffered was broken by the appellants’ own "choices." In this regard it is not arguable that it was their choice to stay in Turkey in a rented house for some time, and then to abandon it and go to Iraq. On the contrary the damage they suffered was a direct result of Frontex’s failures. If the Agency had properly checked the relevant material the appellants would not have been wrongful returned to Turkey and incurred the harm they suffered.

The appellants contend that the harm they suffered was a foreseeable result of Frontex’s failure to fulfil its obligations to uphold fundamental rights during return operations and to prevent individuals from being disembarked in violation of the non-refoulement principle. At the hearing, the question was raised as to whether it was foreseeable that the family, out of fear, would not remain in South-East Turkey but instead flee to Iraq. It has been convincingly argued here  ‘it appears evident that people forcibly expelled to Turkey in 2016 could have suffered chain (i.e., indirect) refoulement to their countries of origin.’  In this case, the damage related to Iraq did not result from an unpredictable intervening event; rather, it was arguably reasonably foreseeable that the Syrian-Kurdish family out of fear to be sent back to Syria by the Turkish authorities would flee to Iraq. If the Court agrees with this, the damage related to Iraq was a direct and foreseeable consequence of Frontex’s unlawful conduct and the GC erred in law by failing to establish this.

Conclusion

The Grand Chamber’s ruling will be highly significant in defining Frontex’s role in protecting fundamental rights during return operations. Judicial scrutiny will enhance accountability and drive further reform, even if the Court must limit its ruling to the specific circumstances of this case. Both member states and Frontex have important obligations to meet when they conduct joint return operations. In this respect, as Eleanor Sharpston, Counsel for the appellants, stated in her concluding remarks: Frontex provides “a safety net that, […] preserves the honour of the European Union as an entity that values and respects the rule of law and fundamental rights.”

 

The Italy/Albania asylum treaty reaches the CJEU: what are the issues?

 

Professor Steve Peers, Royal Holloway University of London

Photo credit: Pudelek, via Wikimedia Commons

 

Is the Italy/Albania deal on housing asylum-seekers a model of the future, to be emulated by the rest of the EU, and the UK – or is it destined to be an expensive failure, like the UK’s Rwanda policy? The fate of the treaty may rest on its compatibility with EU law, and tomorrow’s hearing at the CJEU – in which it is asked to answer several questions from an Italian court on this point – will be the first time that the EU court will have the opportunity to address some of the issues arising from the treaty directly.  What are the issues that the court will address? How might it address them? And what are the possible implications going forward?

Background

As explained in the summary of the national court’s proceedings, the Italy/Albania deal entails moving to Albania some of the asylum-seekers whom Italy at first sight believes to come from ‘safe countries of origin’. The asylum-seekers stay in Albania while their claim for asylum is processed by Italy, applying Italian law – including EU law, which defines the ‘safe countries of origin’ concept. If they win their cases, they are accepted into Italy; if they lose, Italy attempts to remove them.

The CJEU has only given one prior ruling on the substance of the ‘safe country of origin’ concept – its ruling last October on the Czech government’s designation of part of Moldova as ‘safe’. In my discussion of that judgment I summarised the rules on ‘safe country of origin’ in the original 2005 asylum procedures directive, the 2013 asylum procedures Directive (which currently applies to asylum applications), and the 2024 asylum procedures Regulation, which forms part of the recent asylum pact (which will first apply to applications made from July 2026). The definition of ‘safe country of origin’ is more fully laid out there; this blog post focusses on the specific issues directly relevant to the challenges to the Italy/Albania treaty. (It should be noted, though, that the designation creates a rebuttable presumption of safety, which the asylum-seeker can challenge in their particular circumstances; but the application may be fast-tracked and the suspensive effect of an appeal may be limited).  

In the October judgment, the court ruled that it breached EU law to designate part of Moldova as a ‘safe country of origin’, because the current 2013 procedures Directive – unlike the previous Directive or the Regulation applying in future – does not provide for only part of a country to be designated as ‘safe’. The judgment also clarified some elements of the substance of designating a country as ‘safe’ (namely, whether the supposedly ‘safe’ country invoking a derogation from the ECHR always prevented such a designation; the court ruled that it did not, that derogation only being a factor to consider), and on the procedure of challenging such designations in a national court.

Although the October judgment did not directly concern the Italy/Albania treaty, it had an indirect impact on that treaty, which was coincidentally about to be implemented for the first time in practice. Because of the CJEU judgment, Italian courts ruled against detaining asylum seekers in Albania, because Italian law, like Czech law, had also designated some non-EU countries as ‘safe countries of origin’ in part only. Attempting to save the policy, the Italian government adopted a new list of ‘safe countries of origin’, now eschewing the partial designation of any countries. But this list was challenged in turn, and multiple Italian courts have prevented further detentions in Albania, sending a number of questions to the CJEU about the interpretation of other aspects of the ‘safe country of origin’ rules in EU law. In the meantime, the Albanian detention centres, built at great cost to the Italian taxpayer, lie empty except for sheltering stray dogs – to the fury of the Italian government and a confused Elon Musk. But might the doggos nevertheless someday have to make way for human detainees after all?  

The cases before the CJEU

Among a number of cases sent from the Italian courts, the CJEU chose to fast-track two: the ‘Alace’ and ‘Canpelli’ cases. These are fake names invented by the court (applying its usual policy): the former is from Bangladesh, but ‘Alace’ is probably the least Bengali name ever; while the latter sounds like a posh brand of pasta. Presumably the Court’s judgment in these cases will be relevant to answering the questions in the other pending cases (see, for instance, Ortera, Mibone, Capurteli and Leusi), and more broadly determining if the Italy/Albania treaty can still be implemented, and if so, whether any limits may apply.

There are four questions for the Court to answer in Alace and Canpelli. First, does EU law preclude a national legislature from designating a country as a ‘safe country of origin’ itself? Secondly, does EU law require the sources taken into account when designating a country as ‘safe’ to be ‘accessible and verifiable’, as not doing so may complicate a legal challenge to the designation? Thirdly, can the courts use information from sources other than those referred to in the directive (namely ‘other Member States, [the EU asylum agency], UNHCR, the Council of Europe and other relevant international organisations’), when reviewing the designation of ‘safety’? Finally, can a Member State designate a country as ‘safe’ where there are groups of people who are not safe there (for instance, LGBT asylum-seekers, who are specifically mentioned in some of the other Italian court references to the CJEU)?

We can never be absolutely certain how a court will rule, but my answers to those questions (I’m not predicting that the CJEU will agree) would be as follows: there’s nothing in the Directive to prevent a national legislature from designating a ‘safe country of origin’; the sources used when making the designation must be public, to ensure effective judicial review; the list of sources which can be used to make the designation are non-exhaustive (‘including in particular’), so it should follow by analogy that the courts can also use other sources for judicial review of the designation; and under the current law, a country of origin must be safe for everyone before it can be designated as such. The latter should follow by analogy from the Court’s previous judgment on listing part of a country as ‘safe country of origin’, which took account of the legislative history: the prospect of designating a country of origin as ‘safe’ in part was explicitly provided for in the previous Directive and the Regulation applying in future, but not in the current procedures Directive, which suggested that its omission from the current law was deliberate. Likewise the prospect of designating a country of origin as ‘safe’ for some groups was also explicitly provided for in the previous Directive and the Regulation applying in future, but not in the current procedures Directive.

Impact of the future judgment

We can’t fully assess the impact of a future judgment before it is handed down. But while it is possible that the future judgment may criticise aspects of how Italy applies the ‘safe country of origin’ rule, it might still nevertheless be possible for Italy to correct those errors and still apply at least some amended version of the rule to asylum seekers that it wishes to remove to Albania during the processing of their claims. If necessary, the EU might decide to consider bringing forward the application of parts of the future Regulation, or quickly using its power to pass an emergency asylum law.

There might, of course, be further challenges to any amended Italian law, or to EU law itself. For instance, any attempt to provide in EU law that national designations of ‘safe country of origin’ cannot be judicially reviewed, or to provide for hurdles limiting such judicial review, would arguably be invalid as a breach of Article 47 of the EU Charter of Fundamental Rights, which provides for the right to an effective remedy.

So there’s a decent probability that some form of Italy/Albania treaty will be applied in practice – albeit perhaps not as widely as the Italian government hopes. A more significant barrier to the proliferation of such treaties in future may be the reluctance of non-EU countries to agree to them (cf, in a different context, the CJEU judgment ruling that Turkey’s refusal to readmit asylum seekers prevented Greece from deciding that asylum applications made by the asylum seekers in question are inadmissible on the grounds that Turkey is a ‘safe third country’), or at least the limited capacity of such countries (cf Rwanda’s capacity to take only a small proportion of those crossing to the UK on ‘small boats’).

The right to perform secondary movements under international refugee law

 

Paolo Biondi

 

Photo credit: Mstyslav Chernov, via Wikimedia Commons

 

 

Introduction

 

The 1951 Refugee Convention is a pivotal document in international refugee protection, outlining the rights of asylum seekers and the responsibilities of states. Central to this convention is Article 31(1), which addresses the non-penalization of refugees for illegal entry or presence:

 

1. The Contracting States shall not impose penalties, on account of their illegal entry or presence, on refugees who, coming directly from a territory where their life or freedom was threatened in the sense of Article 1, enter or are present in their territory without authorization, provided they present themselves without delay to the authorities and show good cause for their illegal entry or presence

 

Traditionally viewed as a provision ensuring refugees are not penalised for illegal entry or presence if they present themselves promptly to authorities and have good cause for their illegal entry or presence, in my opinion Article 31(1) has broader implications. This blog post explores these implications, highlighting how Article 31(1) and related provisions within International Refugee Law (IRL) and International Human Rights Law (IHRL) should be interpreted to uphold human rights standards. By examining the negotiating history, authoritative interpretations and jurisprudential decisions, it suggests an unorthodox interpretation of Article 31 pointing to a wider interpretation of the right to perform secondary movements where effective compliance with human rights is essential for legitimately allocating of responsibility for asylum applications.

 

Understanding the Orthodox Interpretation of Article 31

 

Article 31(1) aims to address the practical realities of flight from persecution, acknowledging that refugees may have no choice but to enter or remain in a country illegally but, most importantly, implicitly regulates the right to perform secondary movements. During the drafting process, states, particularly France, expressed concerns about ensuring that refugees who had already found protection in one country would not move freely to others without adherence to border formalities. This debate underscored the tension between the humanitarian protection of refugees and states' sovereign rights to control their borders and regulate the distribution of refugees (responsibility sharing).

 

While Article 31(1) prohibits the imposition of penalties on refugees for their illegal entry or presence, it must be read alongside Article 32, which restricts protection from expulsion to refugees lawfully in the territory. This distinction highlights that the scope of penalties under Article 31(1) does not necessarily encompass removal to another country, as this could fall under the separate regime of expulsion under Article 32. Recent jurisprudence, such as rulings on the UK's Rwanda policy, suggests that removal to a third country may not always constitute a penalty under Article 31(1), provided the destination country is deemed capable of providing effective protection. However, international and regional jurisprudence also supports the view that measures resulting in effective penalties, such as removal without substantive guarantees of protection or the risk of refoulement, fall within the ambit of Article 31(1). The principle of non-penalization ensures that refugees are not disadvantaged due to the manner of their arrival, particularly when they lack viable alternatives to reach safety.

 

Two critical elements in the interpretation of Article 31(1) are the concepts of effective protection and meaningful connection. Effective protection refers to conditions in a country that ensure refugees' safety and access to basic human rights. Meaningful connection (e.g. family links) emphasize the importance of a reasonable connection with the transit or destination country. Effective protection is further elucidated in Article 33(1) of the Refugee Convention, which prohibits refoulement—returning refugees to territories where their life or freedom would be threatened. This principle extends beyond the country of origin to any territory where such risks exist, underscoring the necessity of ensuring that refugees are not merely safe but can live free from fear of persecution or other serious harm.

 

During the 1951 Convention negotiations, while Norway’s proposed amendment of Article 31(1)—replacing "coming directly" from a country of persecution with "any territory" where life or freedom is threatened—was not accepted, this rejection does not negate the principle's evolution. Subsequent jurisprudence and interpretations, such as those from the ECtHR, CJEU and the UNHCR, have expanded the understanding of effective protection (e.g. Dublin jurisprudence further below). These developments in essence reflect an acknowledgment that threats in transit countries or deficiencies in protection standards can also justify secondary movements. Yet, already during the negotiations. UNHCR stressed that transit and secondary movements are sometimes unavoidable due to the challenges refugees face in initial asylum countries. Historical examples, such as the flight of refugees during World War II informed this perspective. The experience of UNHCR officials (including the United Nations High Commissioner for Refugees Dr Van Heuven Goedhart) participating in the Conference of Plenipotentiaries on the Status of Refugees and Stateless Persons, who themselves had to transit through multiple countries during their own escapes, highlighted the practical realities faced by refugees.

 

Thus, when refugees seek effective protection or a meaningful connection in a transit or destination country, such as family reunification among other compelling reasons, they may indeed have "good causes" for secondary movements under a modern interpretation of Article 31(1). This aligns with the broader human rights framework that underpins the Refugee Convention, ensuring that its protections remain relevant to contemporary challenges. While the effective protection and meaningful connection terms do not appear explicitly in the text of Article 31, they are crucial because they underpin the broader principles and objectives of the 1951 Refugee Convention.

 

 

 

Secondary Movements and Effective Protection: An Evolving Legal Framework

 

This section argues for a broader interpretation of "good cause" under Article 31(1) and the principle of non-refoulement under Article 33(1) of the 1951 Refugee Convention. It emphasizes that secondary movements can be justified by systemic deficiencies in transit countries, a lack of effective protection, or strong connections, such as family ties, with a destination country. These interpretations align with the Convention's humanitarian purpose, which prioritizes refugees' fundamental rights over rigid procedural constraints.

 

The criterion of "good cause" for illegal entry under Article 31(1) has been a point of contention. Traditional interpretations often limit "good cause" to the manner of entry, focusing on immediate justifications such as the urgency of escape from persecution. However, scholars like Goodwin-Gill advocate for a broader understanding of this concept. According to Goodwin-Gill, "good cause" can extend to reasons such as the presence of family members in a specific country or other genuine links that justify secondary movements. His interpretation aligns with the principle that refugees should not be penalized for seeking effective protection or family reunification, even when this necessitates onward travel. This broader interpretation underscores the humanitarian objectives of the 1951 Refugee Convention, emphasizing that "good cause" should accommodate the complex realities refugees face when fleeing persecution and seeking protection.

 

UNHCR and recent international court rulings have reinforced this expansive view of "good cause" under Article 31(1). For instance, systemic deficiencies in asylum procedures in transit countries—such as those identified in Greece in 2009 during the M.S.S. v. Belgium and Greece case, and in Italy in 2011 with Tarakhel v. Switzerland — demonstrate how inadequate protection can justify secondary movements. These deficiencies are considered violations of human rights, as they fail to meet the minimum standards of effective protection, thereby necessitating the right to seek protection elsewhere. Additionally, the interpretation of non-refoulement has evolved to encompass broader threats beyond persecution as narrowly defined in the 1951 Refugee Convention. This includes situations of generalized violence, systemic oppression, and serious human rights violations, as recognized in both international and regional jurisprudence. Such an expansive approach ensures that refugees are shielded from threats to their life or freedom, whether arising from persecution, armed conflict, or other severe risks, thereby reinforcing the principles of protection enshrined in the Convention.

 

A similar interpretation was upheld in 2014 by the German Federal Constitutional Court (GFCC), which considered the interaction between safe third country practices and Article 31(1). In that context, the court, referencing M.S.S. v. Belgium and Greece, concluded that while Greece was designated as a "safe country" under German law, systemic deficiencies in its asylum procedures rendered it unsafe at the time of entry. These deficiencies exposed applicants to risks of inhumane and degrading treatment, justifying their secondary movements. This interpretation aligns with the reasoning in subsequent decisions by the CJEU, such as in Ibrahim, where the principle of effective protection under the Dublin Regulation was applied to recognize systemic deficiencies as undermining safety. These cases confirm that the right to perform secondary movements is not necessarily linked to "persecution" in a third country; instead, deficiencies in asylum procedures can represent "good cause" and a human rights violation, entitling the applicant to seek protection elsewhere.

 

The absence of protection in third countries fundamentally alters the nature of an asylum seekers journey. What may initially appear as an indirect flight can transform into a direct flight under Article 31(1) when inadequate protection in a transit country prolongs or exacerbates the risk of persecution or other serious harm. This interpretation reflects a broader understanding of the Convention’s purpose: either (a) "good cause" can trump the "coming directly" requirement, encompassing the need to avoid a lack of effective protection en route, or (b) "coming directly" must itself be understood in light of the Convention’s object and purpose, which would exclude transit through countries where effective protection is not available.

 

Such an interpretation aligns with universal rules of treaty interpretation under the Vienna Convention on the Law of Treaties, which emphasize that the object and purpose of a treaty must guide its application. Moreno Lax has argued that unilateral designations of responsibility under the 1951 Refugee Convention or the Dublin system for the application of the Safe Third Country (STC) concept “must be rejected when they merely provide for a deflection mechanism incompatible with the realization of the object and purpose of the Refugee Convention.” While Moreno Lax's critique primarily addresses procedural responsibility mechanisms, the underlying principle she highlights—the necessity of ensuring mechanisms serve the Convention’s purpose rather than undermining it—is adapted here to support the argument against penalizing refugees for secondary movements where effective protection is absent.

 

Similarly, Lord Justice Simon Brown, in R. v. Uxbridge Magistrates’ Court, acknowledged that differing state responses to asylum requests create a rational basis for refugees to exercise some choice in where to seek asylum. While he rejected the idea of unconditional freedom to choose the country of asylum—a stance that might align with concerns over "forum shopping"—his judgment also recognized that there is no obligation under the Refugee Convention requiring refugees to apply for asylum in the first country they enter. This interpretation underscores that the Convention does not preclude secondary movements, especially when linked to legitimate causes such as seeking effective protection or family reunification.

 

Effective protection within the context of Articles 31(1) and 33(1) is crucial. The concept revolves around ensuring that a state can offer genuine safety and uphold the basic human rights of refugees. This includes protection from refoulement, access to fair asylum procedures, and the absence of persecution or inhumane treatment. The Safe Third Country (STC) and First Country of Asylum (FCA) concepts are often invoked in discussions about refugee movements. However, the distinction between these two is largely theoretical. Both concepts require that the third country or the first country of asylum offers effective protection. This means that in the third country concerned, refugees must not be subjected to refoulement or face threats to their life or freedom. The adequacy of protection in the third country must be assessed to ensure compliance with these core principles, ensuring that refugees are not transferred to situations where their fundamental rights may be at risk.

 

Commentator Paul Weis, analyzing the travaux préparatoires of the 1951 Refugee Convention and Article 33(1), concluded that the words "in any manner whatsoever" indicate that Article 33(1) applies to non-admittance at the frontier. This interpretation underscores that an asylum seeker can claim protection under Article 33(1) if they fear persecution or if their physical safety or freedom is endangered in a country where they previously stayed, even before crossing the border. The non-refoulement applies universally, prohibiting rejection at the border in such cases.

 

This understanding is further supported by the absence in Article 33 of any requirement for lawful presence or explicit reference to illegal entry, distinguishing it from other provisions of the Convention, such as Article 31. The absence of such conditions highlights that the protection against refoulement applies broadly to all refugees, regardless of their mode of entry or presence. Other commentators such as Chetail agree with Weis, suggesting that protection against refoulement, grounded in positive obligations, may derive from the principle of non-refoulement itself. This principle obliges states to admit persons at immediate risk of fundamental rights violations, ensuring that non-admission does not have the same effect as refoulement.

 

While the 1951 Refugee Convention does not explicitly regulate access to asylum procedures or territory, the principle of non-refoulement inherently requires that refugees not be returned to situations where their life or freedom would be threatened. The refugee definition applies specifically to individuals outside their country of nationality. In contrast, IHRL imposes no such geographical limitation, and non-refoulement, from a human rights perspective, applies to any person under another state's effective control, regardless of location. While IRL and IHRL operate as distinct legal regimes, their objectives often intersect. For example, individuals may be protected from expulsion under human rights law without necessarily qualifying as refugees under the 1951 Refugee Convention. The extraterritorial application of the non-refoulement principle, recognized in both IRL and IHRL, extends states' obligations to prevent the return of individuals to harm, even when they are not within the state’s territory. This principle has been extensively developed in case law, particularly by the European Court of Human Rights (ECtHR) in cases such as Hirsi Jamaa v. Italy, where the Court confirmed that effective control triggers non-refoulement obligations.

 

 

Balancing 'Good Cause' and 'Coming Directly' Under Article 31(1)

 

This section expresses reservations about the 2017 UNHCR interpretation of "good cause" under Article 31(1) as narrowly focused on the nature of illegal entry, arguing that this view is inconsistent with the provision's negotiating history and broader judicial interpretations.

 

In 2017, UNHCR published a research paper that interprets the "good cause" criterion in Article 31(1) as primarily relating to the nature of illegal entry—specifically, the need to demonstrate a "good cause" for entering a country illegally. The publication argues that "the good cause requirement should not be used to rehearse arguments relating to safe third countries as these matters are examined under the ‘coming directly’ element." While this interpretation is plausible, it appears inconsistent with the negotiating history of Article 31(1) and with interpretations from international courts.

 

Authoritative figures, such as Goodwin-Gill, have challenged the conventional understanding of "good cause" in the context of irregular entry under Article 31(1). Goodwin-Gill argues that the "good causes" criterion should not be confined solely to the manner of illegal entry. Instead, it may encompass a broader range of reasons for seeking asylum in a particular country, including family reunification or systemic deficiencies in transit countries. These reasons reflect the Convention’s humanitarian purpose and recognize the complex realities refugees face during their flight to safety.

 

The relationship between "good cause" and "coming directly" has been a matter of debate. One interpretation is that these are separate, non-overlapping conditions, meaning "good cause" cannot override the "coming directly" requirement. However, according to me scholars like Goodwin-Gill argue that "good cause" should inform the application of "coming directly," allowing exceptions when transit countries fail to provide effective protection or there is no meaningful link. This perspective aligns with the Convention’s humanitarian objectives, ensuring that refugees are not penalized for seeking safety where protection was unavailable en route.

 

While Article 31(1) does not allocate responsibility for asylum applications, its principles shape the treatment of secondary movements. For example, family bonds or other relevant links in alternative destinations are recognized in regional frameworks, such as the EU’s Dublin Regulation, which prioritizes family links, and the "safe third country" concept, which requires a meaningful connection beyond mere transit. In 2020 in the FMS case the CJEU also clarified that mere transit cannot be intended as meaningful  connection for the purpose of STC concept application. These interpretations reinforce the need to assess secondary movements within a legal framework that respects refugees' rights. Most recently another key aspect was clarified in the CJEU judgment in Elliniko Symvoulio where it is established that if the third country designated as generally safe by a Member State does not in fact admit or readmit the applicants for international protection concerned, that Member State cannot reject their applications for international protection as inadmissible on the basis of Article 33(2)(c) of the Asylum Procedures Directive (APD).

 

Judicial decisions, such as FMS, provide critical clarification on the application of "good cause" and "coming directly," supporting broader interpretations that prioritize refugee rights and systemic fairness. Such decisions are all the more important during a time EU Member States are exploring the possibility to remove the meaningful connection requirement from the newly adopted Asylum Procedure Regulation. In the same vein, Prof. Steve Peers critiques the revised Dublin rules for narrowing the scope and effectiveness of appeal rights, which, as he notes, marks a shift 'towards a purely intergovernmental framework' rather than one that individuals can effectively engage with to enforce their rights. This procedural limitation highlights the systemic barriers refugees face in seeking fair assessments, particularly in the context of systemic deficiencies in transit countries. His critique underscores the necessity of maintaining procedural safeguards to ensure fairness and uphold refugees' rights under Article 31(1). The FMS ruling aligns with the notion of "good cause" under Article 31(1), which advocates for flexibility and fairness in determining when secondary movements are justified. It supports the argument that a lack of substantive ties in a transit country may justify onward movement. The FMS case underscores the importance of individual assessments and procedural safeguards. This aligns with the broader goal of ensuring that Article 31(1) is applied in a way that prioritizes fairness and refugees' rights. While FMS primarily interprets EU law, it indirectly supports a broader understanding of "good cause" and "coming directly" by emphasizing meaningful connections and rejecting procedural shortcuts like reliance on mere transit.

 

 

 

Protecting Family Unity in the Context of APD and Dublin Regulation Interplay

 

Now, we turn to the interplay between EU asylum law and IRL, particularly how they address family unity, meaningful connections, and procedural safeguards in allocating responsibility for asylum seekers. The EU asylum system closely reflects the principles of IRL, especially those enshrined in the 1951 Refugee Convention. This alignment provides a valuable framework for examining how EU rules incorporate and interpret these international standards, with a particular focus on family links and responsibility allocation.

 

Article 38(2)(a) of APD stipulates that the application of the safe third country concept must consider a connection between the asylum seeker and the third country. This connection, which could include family ties, must be reasonable for the applicant to relocate to that country. The directive itself explicitly states that the presumption of safety in a third country can be rebutted by the applicant, emphasizing the importance of individual assessments to ensure access to asylum procedures. Misapplication of admissibility rules could unjustly deny access to appropriate asylum procedures, thereby impacting the applicant's fundamental rights.

 

The Dublin Regulation, referenced by Article 38 of the APD, assigns responsibility for examining asylum applications through a comparative test. This test prioritizes family links and other meaningful connections to determine the most suitable EU Member State to handle the asylum claim. However, the regulation also considers other factors, such as the first EU state of entry, which can play a significant role in responsibility allocation. According to EU law, family unity remains a crucial factor, and decisions should aim to preserve these ties wherever possible.

 

The application of the APD procedure, particularly the "safe third country" rule under Article 33(1), has been interpreted differently in recent case law. On 17 March 2016, the CJEU in the Mirza case addressed the interplay between the STC rule under Article 33 of the recast APD and the Dublin Regulation. The Court ruled that a Member State may apply the STC rule to declare an application inadmissible, even if it is not the Member State responsible for examining the claim under the Dublin Regulation. This judgment reflects an interpretation of the Dublin Regulation’s explicit provisions regarding STC rather than a general approach to inadmissibility.

 

While the Mirza judgment interprets EU law as it stands, in my view concerns remain about its implications for procedural fairness and fundamental rights. For example, the judgment permits Member States to invoke the STC rule without fully determining responsibility under the Dublin Regulation, which could undermine the structured allocation of responsibility designed to safeguard family unity and other meaningful connections. EXCOM Conclusion 15 emphasizes that asylum should first be requested in a state where the applicant has a connection, such as family ties, and this requires a comparative test, not a non-contextual application of inadmissibility rules. Similarly, Article 31(1) of the 1951 Refugee Convention recognizes family links as a "good cause" for onward movements, highlighting that strict and non-comparative procedures risk penalizing refugees contrary to international law.

 

In my opinion, the Mirza judgment illustrates a tension between procedural flexibility for Member States and the need for structured, rights-based responsibility allocation. While the judgment aligns with EU law’s textual framework, the absence of a comparative test in applying the STC rule risks decisions that fail to account for family unity or systemic deficiencies in protection. To strengthen procedural safeguards, inadmissibility decisions under Article 33(1) APD should ideally follow responsibility determinations under Article 3 of the Dublin Regulation. Although this sequencing is not explicitly required by the regulation, it reflects a normative approach that aligns better with the principles of family unity and effective protection embedded in both EU law (Fundamental Rights Charter) and IRL.

 

This cross-referencing aims not to override the Dublin responsibility criteria, principally to safeguard family unity (Art. 8 DRIII) criteria or the humanitarian clause (Art. 17 DRIII), which can and should be used to ensure family unity once a person gains access to the EU. The ECtHR has affirmed the proactive obligations under Article 8 ECHR in cases such as Senigo Longue and Tanda-Muzinga, which must be read in conjunction with the primacy of the Best Interests of the Child and the respect of family unity. This reiterates the positive duties of states to proactively and expeditiously comply with the procedural and substantive elements of Article 8 ECHR, both as an autonomous right and as part of an instrument of secondary legislation. Doing otherwise would undermine family unity (including for dependency reasons), with inadmissibility taking precedence over it.

 

Currently, it is possible for the principle of non-penalization under Article 31(1) of the 1951 Refugee Convention to be undermined in the EU context. Article 38(1)(e) of the recast APD allows an applicant to be sent to a third country outside the EU where there is a "possibility" to request refugee status. However, the directive requires that a connection between the applicant and the third country be established, as outlined in Article 38(2)(a). Case law further clarifies that mere transit through a third country is insufficient to meet this requirement, emphasizing the need for a meaningful connection.

 

Additionally, the Dublin Regulation governs responsibility allocation for examining asylum claims among EU Member States and does not directly apply to transfers to third countries under the APD. While the Dublin Regulation does not explicitly mandate a "comparative test," its criteria prioritize factors such as family unity and the applicant’s meaningful links to a Member State. A rigorous application of these principles ensures that inadmissibility decisions under Article 33(1) of the APD do not override the safeguards enshrined in the Dublin Regulation and EU asylum law as a whole.

 

 

Conclusion: Balancing Control and Protection

 

Article 31(1) of the 1951 Refugee Convention does not prohibit the diversion of asylum applications to third countries willing and able to provide effective protection, even in the absence of significant links to those countries. However, EXCOM Conclusion 15 and UNHCR guidelines emphasize that the principle of non-penalization under Article 31(1) limits such transfers when meaningful connections or effective protection are lacking. Justifications for secondary movements include insufficient protection in transit countries, refusal of protection, or strong family links in another state, which are recognized as "good cause" for onward movement.

 

A comparative test is necessary to evaluate the applicant’s connections to third countries versus their ties to other states, including within the EU under the Dublin Regulation. Effective protection, meaningful links, and family unity extend beyond non-refoulement, shaping a state's discretion in allocating responsibility for asylum claims. While Article 31(1) does not establish a positive obligation to admit or always justify secondary movements, it imposes constraints on measures that ignore these factors, as such restrictions may constitute penalties or harm integration prospects.

 

The interpretation of Article 31(1) requires a balance between state sovereignty and refugee protection. States have the right to control borders but must also honour international obligations, ensuring access to effective protection and family unity. Refugees should not face penalties for irregular entry or secondary movements driven by genuine humanitarian reasons. By applying Article 31(1) with sensitivity to these realities, states can uphold the 1951 Refugee Convention’s principles while maintaining humane and just asylum policies.

AI Literacy Under the AI Act: An Assessment of its Scope

  

Tiago Sérgio Cabral*

** PhD Candidate at the University of Minho | Researcher at JusGov | Project Expert for the Portuguese team in the "European Network on Digitalization and E-governance" (ENDE). Author’s opinions are his own.

Photo credit: Salino01, via Wikimedia Commons

 

Introduction to AI Literacy under the AI Act

 

Under Article 4 of the AI Act (headed “AI literacy”), providers and deployers are required to take “measures to ensure, to their best extent, a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf, taking into account their technical knowledge, experience, education and training and the context the AI systems are to be used in, and considering the persons or groups of persons on whom the AI systems are to be used”. The concept of AI literacy is defined in Article 3(56) of the AI Act as we will see below.

Article 4 is part of a wider effort by the AI Act to promote AI literacy, which is also reflected in other provisions such as those addressing human oversight, the requirement to draw up technical documentation or the right to explanation of individual decision-making. Article 4 focuses on staff and people involved in the operation and use of AI systems. As such, the main consequence of this provision is that providers and deployers are required to provide training for employees and people involved in the operation and use of AI systems, allowing them to obtain a reasonable level of understanding of the AI systems used within the organization, as well as general knowledge about the benefits and dangers of AI.

It is important to distinguish between the rules on AI literacy and the requirements on human oversight, in particular Article 26(2) of the AI Act. Under this provision, deployers will be required to assign human oversight of high-risk AI systems to natural persons who have the necessary competence, training and authority, as well as the necessary support. The level of knowledge and understanding of the AI system required of the human overseer will be deeper and more specialized than what is required from all staff in the context of AI literacy. The human overseer must have specialized knowledge about the system that he/she is overseeing. The people subjected to AI literacy obligations require more general knowledge about the AI systems used in the organization, particularly the ones with whom the staff is engaging, along with understanding of the benefits and dangers of AI. The level of AI literacy required in organizations that develop or deploy high-risk systems will, naturally, higher than organizations that deploy, for example, systems subject to specific transparency requirements. In any case, it will still likely be lower than one is required of a human overseer (although not limited to high-risk systems as the rules for human oversight).

 

The scope of Article 4 of the AI Act

 

AI literacy is a sui generis obligation under the AI Act. It is systematically placed within “Chapter I – General Provisions”, and thus disconnected from the risk categorization for AI systems. This can result in significant challenges in interpreting the scope of the obligations arising from Article 4 of the AI Act.

In fact, an isolated reading of this provision could result in the conclusion that AI literacy obligations apply to all systems that meet the definition of an AI system under Article 3(1) of the AI Act. As the definition in Article 3(1) of the AI Act is extremely broad in nature, this would result in a large expansion of the scope of the AI Act, far beyond the traditional pyramid composed of (i) prohibited AI (Article 5 of the AI Act), (ii) high-risk AI systems (Article 6 of the AI Act); and (iii) AI systems subject to specific transparency requirements (Article 50 of the AI Act). The risk categorization also includes general-purpose AI models and general-purpose AI models with systemic risk, but the AI literacy obligation under Article 4 appears to only apply directly to systems. Indirectly, providers of AI models will be required to provide providers of AI systems that integrate their models with sufficient information to allow the latter to fulfil their literacy obligations (see, inter alia, Article 53(1)(b) of the AI Act).  

The abovementioned interpretation does not hold, however, if we opt for reading of Article 4 of the AI Act that adequately considers the definition of AI literacy under Article 3(56) of the AI Act. Article 3(56) of the AI Act lays down that AI literacy means the “skills, knowledge and understanding that allow providers, deployers and affected persons, taking into account their respective rights and obligations in the context of [the AI Act], to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause”.

Providers and deployers of AI systems that are not part of the abovementioned categorization are not, strictly speaking, subject to any obligations related to these AI systems– at least none arising from the AI Act. Likewise, the AI Act does not grant affected persons any rights in relation to systems that are not part of the “classic” categorizations. Given that the existence of rights and obligations are the building blocs upon which the AI literacy measures need to be designed, if they do not exist, the logical conclusion is that the definition cannot be applied in this context. If the definition under Article 3(56) of the AI Act cannot be applied, Article 4 of the AI Act which entirely depends on this definition cannot apply either.

 

Enforcement

 

In addition to issues around the interpretation of its scope, enforcement of Article 4 also raises significant questions. Article 99(3-5) of the AI Act does not establish fines for the infringement of AI literacy. As such organizations cannot be fined for failing to fulfil their AI literacy obligations based on the AI Act (if considered in isolation). Market surveillance authorities have enforcement powers that do not entail financial sanctions, but it is still a strange scenario for the AI Act to establish an obligation without a corresponding fine which is arguably the key sanctioning tool. It also remains to be seen whether market surveillance authorities will prioritize an obligation that the EU legislator did not consider as significant enough to merit inclusion in Article 99(3-5) of the AI Act.

In addition, Member States may use their power under Article 99(1) of the AI Act to establish additional penalties and, through those, ensure the enforcement of Article 4 of the AI Act. However, this approach risks fragmentation and inconsistency, which is undesirable.

Private enforcement is also a possibility, but whether in the context of tort liability or product liability, it seems to us that proving damages and the causal link between the behaviour of the AI system and the damage may continue to be major obstacles to the success of any attempts. In this context, it is important to note that new EU Product Liability Directive (applicable to products marketed or put into service after 9 December 2026) contains relevant provisions that may make private enforcement easier against producers in the future. In particular, Article 10(3) of the Product Liability Directive establishes that “the causal link between the defectiveness of the product and the damage shall be presumed where it has been established that the product is defective and that the damage caused is of a kind typically consistent with the defect in question”. In addition, Article 10(4) addresses situations where claimants face excessive difficulties, in particular due to technical or scientific complexity, in proving the defectiveness of the product or the causal link between its defectiveness and the damage by allowing courts to establish a presumption. However, in this scenario, linking a breach of the obligation to ensure AI literacy to a defect in a product or a specific instance of damage in any type of satisfactory manner seems challenging and unlikely to accepted by courts.

Lastly, although the AI literacy obligations technically became applicable on 2 February 2025, the deadline for the appointment of Member State authorities is on 2 August 2025. As such, any attempt of enforcement will likely be limited during this period.

 

Identification of AI systems as a preliminary step for the assessment of literacy obligations

 

Although, as referred above, literacy obligations are not applicable to systems outside of the risk categorization of the AI Act, from an accountability perspective, providers and deployers who want to rely on this exception should still proceed an evaluation of the AI systems for which they are responsible as a preliminary step. Only after identifying the AI systems and evaluating whether they fall out of the risk categories established by the AI Act can providers and deployers know with an adequate level of certainty that they are not likewise subject to the literacy obligations under Article 4 of the AI Act.   

 

The GDPR as an alternative source of literacy obligations

 

For providers and deployers who are acting as data controllers under the GDPR, it is important to note that non-applicability of Article 4 of the AI Act does not exclude literacy and training obligations that may arise under other EU legal instruments. Particularly, for AI systems that depend on the processing of personal data for their work, adequate training of staff may be required to comply with controller accountability obligations and ensure that the measures implemented by the controller to ensure lawful processing of personal data in the context of the organization’s use of AI (Articles 5(2) and 24 of the GDPR). Considering the wording of Article 39(1)(b) of the GDPR, data protection officers should be involved in the evaluation of training requirements.