5.4.5. When can knowledge be inferred from the general context rather than established through direct evidence?

Even when purpose-based intent is not required, intent remains inextricably linked to knowledge; there cannot be intent without knowledge. The question addressed in this section is therefore distinct from, but closely connected to, the intent threshold discussed in the previous section. It concerns how far courts are willing to infer the applicant's knowledge from circumstantial evidence rather than requiring direct proof of what the applicant actually knew, and what level of specificity is required before general contextual evidence can be treated as establishing individual awareness.

EUAA guidance provides that knowledge of the specific crime is not required for aiding and abetting insofar as the applicant knew that one of a number of such acts will probably be committed and one of them was in fact committed. It is sufficient that the applicant was aware, based on the circumstances, that their activities would assist or facilitate an excludable act. When the applicant's conduct is more remote from the principal crime, knowledge may need to be inferred from the general circumstances rather than established through direct evidence. Such inference may be made from circumstantial evidence, including information about the general context in which the conduct occurred, the scale of atrocities committed and their general nature in a given region or country, but it should be borne in mind that the applicant may not have possessed this information at the time of the acts. Inferences from COI must therefore be drawn carefully, and the availability of COI does not automatically establish that the applicant had access to it or acted on its basis.

When the abuse is more widespread and systematically documented, such as cases involving crimes against humanity, courts have been more willing to infer knowledge objectively from the general context. This is illustrated by the Syrian security force cases, in which both the Belgian CALL in X v CGRS (16 April 2024) concerning a former soldier who manned a checkpoint and the Dutch District Court of the Hague in Applicant v The Minister for Asylum and Migration (14 November 2024) concerning a sergeant who conducted background checks inferred knowledge not only from the applicants' own statements but from the widespread and systematic nature of the abuses documented in COI. Both courts considered that anyone serving within those structures during the relevant period should have been aware of the crimes being committed.

Nevertheless, even when systematic abuse is documented by COI, an individualised assessment remains necessary to determine whether the specific applicant possessed the requisite knowledge. This is illustrated by two cases involving Afghan applicants who performed comparable roles within the same authority, yet the courts reached opposite conclusions. On 7 April 2025, in Applicant v Finnish Immigration Service (mentioned in Section 5.4.3), the Finnish Supreme Administrative Court excluded an applicant who worked with the Afghan National Directorate of Security (NDS) between 2011-2014, arresting persons who were subsequently handed over to the intelligence unit. Although the applicant claimed he was not certain whether torture was being carried out by the other unit, the court did not accept this. It gave decisive weight to his own statements that he had suspected torture was occurring, based on overhearing conversations about it and observing detained persons who were unable to walk properly. Combined with COI establishing the widespread and systematic nature of torture across the NDS during that period, the court concluded that the applicant must have been aware of what was happening to those he helped to arrest.

A different conclusion was reached by the Danish Refugee Appeals Board in afgh20252 (23 January 2021) concerning a comparable case of an applicant who worked with the NDS from 2017-2021. The applicant claimed he was unaware of torture by other units and had only a general perception that it could occur. The board treated this as insufficient to fulfil the knowledge requirement, accepting his explanation that torture was prohibited by law and that he had been informed of this during training, findings which it corroborated with COI. Critically, rather than relying on evidence of torture across the NDS as a whole, the board examined the COI with considerable specificity, looking at the province where the applicant worked, the time period of his service and the unit to which he belonged. On that more precise reading, the picture was considerably less clear-cut than in the Finnish case. Country reports showed that rates of torture had decreased since 2011 and varied significantly across provinces, with only 7 out of 42 detainees in the applicant's specific province reporting torture between 2017-2018. Taking into account the applicant's subordinate position, his non-participation in interrogations and his lack of knowledge about the working methods of other units, the board concluded that individual awareness of torture could not be inferred from the general context alone.

A more recent decision by the Danish Refugee Appeals Board (Afgh/2025/41/lafal), pronounced in September 2025, also concerned an Afghan applicant who was employed with the NDS but for a longer period of time (1982-1990 and 2010-2013). This case concerned a different time period from the previous cases, as it assessed the applicant’s participation in approximately 350 security operations which involved the capture and handing over of prisoners to other departments from 1985-1990. Based on COI, the board established that during the Afghan-Soviet war, torture was carried out systematically by the NDS, which had interrogation offices across all provinces in Afghanistan. Although the applicant denied having knowledge of the torture being carried out, the board found this unconvincing in light of his rank and long duration of employment and upheld his exclusion from international protection.

The different outcomes reflect not only the different time periods and provincial contexts in which the applicants served, but also a difference in the weight given to the applicants' own statements and the level of specificity at which COI was examined. In the Finnish case, the applicant's own account provided a basis for inferring knowledge that the court treated as decisive. In the first Danish case, the absence of such indicators, combined with a more detailed reading of the COI, pointed in the opposite direction. On the other hand, the second Danish case demonstrates that inference of knowledge cannot be based only on COI and must also be assessed based on the individual circumstances of the applicant, including their rank and length of employment. The cases illustrate both the importance of examining COI at an appropriate level of specificity and the significance of the applicant's own statements and role to the knowledge assessment.

These cases point to a practical challenge for decision-makers and courts. The main difficulty is knowing when the general context is of such a nature to justify concluding that a particular applicant must have been aware of the crimes taking place. The framework permits knowledge to be inferred from the general context but does not give precise guidance on where that threshold lies; is it enough to establish that abuse was widespread across an organisation as a whole, or must the analysis go further and look specifically at the unit, province, time period and role of the individual applicant? The Danish case suggests the latter, but not all courts have applied that level of detail. This uncertainty, combined with the fact that sufficiently detailed COI may not always be available, creates a real risk of divergent outcomes for applicants in broadly comparable situations.

 


The CJEU has been instrumental in embedding the principle that exclusion cannot be applied automatically, holding in B and D and Lounani that an individualised assessment of the applicant's specific role, knowledge, contribution and personal circumstances is always required to establish individual responsibility. This principle has been upheld consistently by national courts and has since been codified in the QR.

Yet as the case law demonstrates, despite this clear direction, national courts continue to diverge in how they apply the assessment framework in practice. Disagreements persist over what constitutes a substantial contribution to an excludable act, what level of intent is required and how far knowledge can be inferred from the broader systemic context. These divergences may produce materially different outcomes for applicants in comparable situations, undermining the consistency that the CJEU's case law was designed to deliver. Greater harmonisation remains essential if the exclusion clauses are to be applied in a manner that is both principled and uniform across the EU.