CX1 & Ors, R (on the application of) v Secretary of State for Defence & Anor [2024] EWHC 892 (Admin)

[2024] EWHC 892 (Admin)Case No AC-2023-LON-001736and AC-2023-LON-002091
IN THE HIGH COURT OF JUSTICE
KING’S BENCH DIVISION
DIVISIONAL COURT
Venue Royal Courts of Justice, Strand, London WC2A 2LLDate 19 April 2024
Before
LORD JUSTICE DINGEMANSVice-President of the King’s Bench DivisionMR JUSTICE JOHNSONMR JUSTICE CHAMBERLAIN
Between
THE KINGClaimant(on the application of CX1, CX2, CX4, CX6 and CX7)ClaimantSECRETARY OF STATE FOR DEFENCEDefendantSECRETARY OF STATE FOR FOREIGN, COMMONWEALTH AND DEVELOPMENT AFFAIRSDefendantAnd between:ClaimantTHE KINGClaimant(on the application of MP1)ClaimantSECRETARY OF STATE FOR DEFENCEDefendantMP2, MP3, MP4 and MP5Interested PartiesMartin Goudie KC, Dominic Lewis and Alex Jamieson (instructed by the Special Advocates Support Office) as Special AdvocatesSir James Eadie KC, David Blundell KC Richard O’Brien KC, Nicholas Chapman, John Bethell, Natasha Jackson and Luke Tattersall (instructed by the Treasury Solicitor) for DefendantsHearing Hearing date: 26 March 2024OPEN version of CLOSED JudgmentThis is an OPEN version of a CLOSED judgment which was handed down on 19 April 2024. It is published on 8 August 2025 following a hearing on 1 August 2025 before Lord Justice Dingemans and Mr Justice Johnson. Some parts of the original judgment have been redacted (removed completely) and some parts have been reworded as gists of the original judgment. The redactions and gists reflect agreements reached between the Special Advocates and counsel for defendants, and have been made for national security purposes. In most, but not all cases, the redactions and gists are indicated by being underlined and within square brackets. We do not exclude the possibility that further parts of the CLOSED judgment may be opened up, and we grant a general liberty to apply for that purpose.Mr Justice Johnson:1.This is the judgment of the court.

Introduction

[2]The underlying OPEN claims for judicial review in these cases each raised issues about the application of the Afghan Relocations and Assistance Policy (ARAP). ARAP governs the circumstances in which His Majesty’s Government will grant leave to relocate to the UK. In each of these cases the claimants challenged decisions that they were not eligible for relocation to the UK under ARAP. OPEN judgments have been handed down in each of these cases: [2024] EWHC 94 (Admin) (the CX cases) and [2024] EWHC 410 (Admin) (the MP cases).[3]For the reasons given in the judgment in the CX cases, the court quashed decisions that CX1 and CX6 did not qualify for relocation to the United Kingdom under ARAP, and remitted their applications to the defendants to be redetermined. CX2’s claim was withdrawn. The claims of CX4 and CX7 were dismissed. Since then, arrangements have been made for CX1 to relocate to Canada. The special advocates agree that it is not necessary to consider his case further.[4]For the reasons given in the judgment in the MP cases, the court quashed the decision that MP1 did not qualify for relocation to the United Kingdom under ARAP. There is an outstanding application for permission to appeal against that decision. MP2, MP3, MP4 and MP5 are family members of MP1 whose claims are contingent on the decision-making in his case.[5]This CLOSED judgment concerns an entirely separate challenge, brought as part of the same proceedings, to a decision that the claimants do not know about but which affects their interests. That decision relates to underlying events which, again, the claimants do not know about. Those underlying events give rise to a risk that they (and many others) might be subject to torture, and might be at risk of death. The challenge has been advanced on their behalf by the special advocates.[6]In each case, the first defendant sought a declaration that a closed material application could be made under the Justice and Security Act 2013 and Part 82 of the Civil Procedure Rules. The matters underpinning the request for such declarations concern issues that are quite distinct from the substantive issues in the OPEN claims and do not bear on the decisions to allow those claims. They arise as a result of the defendants’ duties of candour. Special advocates were appointed. They did not contest the applications, made under section 6 of the 2013 Act, for declarations that closed material applications may be made. They also do not contest the defendants’ applications, made under section 8 of the 2013 Act, for permission not to disclose material otherwise than to the court and the special advocates. On the evidence, and for reasons which will become clear, they were right not to do so. We grant those applications.[7]The closed hearings, and this CLOSED judgment, involve a significant departure from the fundamental principle of open justice. For the reasons we explain below, that departure is justified in the exceptional circumstances of this case. This is, however, something that we will keep under review. It is likely that a time will come (possibly within weeks or months) when a continued departure from the principle of open justice can no longer be justified. If and when that time comes, we will publish this judgment (if necessary, subject to such redactions as are necessary and can be justified).[8]As will be clear, the factual background is exceptionally sensitive. It would, if publicly disclosed now, give rise to a risk to the lives and safety of many people. If and when the time comes where this judgment can be published, it may still be the case that aspects will remain sensitive. For that reason, we have deliberately not set out all of the details of the case where it is not absolutely necessary to do so. That may make it easier to publish this judgment sooner than would otherwise have been the case.[9]The stark background is that a data leak by [Individual 1] acting on behalf of the first defendant has resulted in the lives of almost 100,000 people being put at risk. Many of them live in Afghanistan. The defendants’ case is that public acknowledgement of the data leak (or even telling those affected) would significantly increase the risks. That severely curtails the protective measures that can be put in place, at least until knowledge of the data leak becomes publicly known (referred to by the defendants as a “break glass” event).[10]We acknowledge that the underlying circumstances have given rise to exceptional challenges for the defendants. It is not possible fully to mitigate the risks to so many lives that have been created by the data leak. The defendants have to take account of many different, sometimes competing, considerations when determining the appropriate policy to respond to the data leak and to mitigate the risks that have been created. It is clear from the evidence that substantial consideration has been given to the issues that arise at the most senior official and Ministerial level.[11]An important decision was made by the first defendant the day before the hearing, which substantially advanced earlier decision-making. The special advocates were content to treat that as, in effect, the decision that is now under challenge in both cases. Notwithstanding the general rule that the court does not entertain “rolling judicial review claims” (see Administrative Court Judicial Review Guide para 7.11.1), the first defendant was content, given the unusual circumstances of this case, to respond to that challenge.[12]We are grateful to all counsel, and the wider legal teams, for the way in which this sensitive and concerning case was conducted, including their willingness to litigate the issues at a substantive hearing that took place less than 24 hours after the (effective) decision under challenge. Further delay would have been undesirable. That is partly because of the risks arising from the data leak, and partly because there are other cases waiting to be heard that raise the same issues. The background

The data leak

[13]It was sometimes necessary to check information contained in ARAP applications [Redacted].[14][In February 2022 Individual 1 inadvertently emailed a dataset containing the personal data of around 25,000 ARAP applicants and their family members (totalling approximately 100,000 people)]. These included CX1, CX6 and MP1 (but not CX4 or CX7). The personal data relating to CX1, CX6 and MP1 includes the fact that they have made applications to the ARAP scheme, the date of their applications, their full names, email addresses and phone numbers.[15]On 14 August 2023 it became clear that there had been onward disclosure of the data set to individuals who were not authorised to have it. Extensive work has been undertaken to try to control and contain the dissemination of the data set (including by way of legal proceedings). To some extent this work has been successful (in that the information available to the court is that the data set has not entered the public domain, and nor is there any evidence that it has come to the attention of the Taleban). It is not necessary or desirable to set out the detail of that work in this judgment. Response to the leak[16]Nina Cope, the Chief Operating Officer at the Ministry of Defence (“MoD”), was appointed to lead a group to oversee the response to the data leak. Ministers have also been involved. On 16 August 2023 the Information Commissioner’s Office was informed, and a report was submitted to the ICO the following day. The matter was also referred to the Metropolitan Police. On 17 August 2023, David Williams, a journalist from the Daily Mail approached the MoD and said that he was aware of the data leak and that he intended to run a story about it. He agreed not to do so until after the MoD had undertaken protective work. A member of the public also approached the MoD and said that they were aware of the data leak.[17]On 22 August 2023, Lewis Goodall, a journalist with “The News Agents” podcast, approached the Foreign, Commonwealth and Development Office to indicate that they had information about the data leak. Mr Goodall also agreed not to run the story until protective work had been undertaken by the MoD.[18]The first defendant assessed that the data leak gave rise to a risk to life. In particular, it was highly likely that the Taleban would become aware of the dataset and would then acquire the dataset, and would then exploit the dataset. It was assessed that it was highly likely that the Taleban would prioritise targeting high-profile individuals named in the dataset, [although all individuals on the dataset would face significant risks]. Targeting was assessed to include torture and extra judicial killings.[19]The first defendant sought the exceptional remedy of an injunction, contra mundum, to restrain further publication of the dataset. The claim form states:
“…the Claimant seeks an injunction, initially for a period of four months, to prevent… publication or other disclosure [of the dataset], while the Claimant takes what steps he can to protect those named in the data set.”
[20]The application for the injunction was supported by a witness statement from Ms Cope. In the course of that statement Ms Cope said:
“The Claimant seeks an injunction to prevent such publication while it undertakes urgent targeted protective action to mitigate the risk to life and risk of persecution to the people named in the data. … The Claimant has and will continue to take actions to reduce risk to these individuals as soon as practically possible. … The Claimant cannot give an exact timeline for when this work will be complete, and all reasonable mitigations will have been put in place. The current estimate that it could be in the region of four months.”
[21]On 1 September 2023, Robin Knowles J granted injunctive relief contra mundum to prohibit further dissemination of the data set. The order also prohibited the disclosure of the fact that the order had been made (“a super injunction”). That order has been continued by further orders made by Chamberlain J. Those orders have been made after hearing, in private, representations made on behalf of those journalists who are aware of the data leak and also after hearing, in CLOSED, from special advocates. The reasons for those orders are explained, in detail, in judgments that are necessarily not published (but are likely to be published at some point in the future): [2023] EWHC 2999 (KB) and [2024] EWHC 312 (KB). A further hearing to determine whether the injunction should be continued is listed on 30 April and 1 May 2024.[22]In a witness statement dated 10 October 2023, Deana Rouse, the Deputy Chief Operating Office at the MoD explained the steps that were being taken as a result of the data leak. [Redacted]. In respect of those who were not eligible under ARAP, she said:
“MOD continues to work with other Government departments to develop possible options to assist individuals currently deemed ineligible for relocation under ARAP. This cross-Government work is being informed by the latest threat assessment and outcomes of the impact assessment. The Claimant considers that the continuation of the [super injunction] is vital for the effectiveness of this work.”
[23]On 16 November 2023, officials presented three options to Ministers as to the response that could be made to the data incident. Option A was to ensure that all those identified in the dataset and their family members should have access to a route to the UK. Option B was not to create bespoke access to the UK for any individual identified in the dataset. Option C was to create bespoke access to a route to the UK for a targeted cohort of around 150 individuals who were not eligible under the ARAP scheme but who had been directly affected by the data incident, [had links to His Majesty’s Government and who were in high-profile roles (and were thus said to be at particular risk)].[24]On 16 November 2023, Ministers gave directions to identify a targeted cohort of approximately 150 of those identified in the data set who were at the highest risk, and to prepare plans to relocate them and their family members in the event that the Taleban gained access to the dataset. In effect, this was a direction to work up plans to implement option C. A direction was also given to revert in the event of a significant increase in the size of this targeted cohort. Thereafter, the MoD refined the criteria for the “targeted cohort”, which included that the individuals should have a confirmed link with the UK Government.[25]On 17 November 2023, Natalie Moore, the Director of the Defence Afghan Relocations and Resettlement Directorate, made a witness statement in the CX proceedings. She said:
“All of the Claimants have been deemed ineligible for ARAP. However, only CX1 and [CX6] are included within the database, which is the subject of the data incident, and therefore fall within the “ineligible cohort”
. No relocation action will be taken in respect of those who are ineligible for ARAP and are not within the ineligible cohort. …I have decided that MOD will not take any further relocation action in respect of any of the Claimants. With respect to CX1 and [CX6], it is expected that they would not be eligible for the targeted relocation plan directed by Ministers, based on the proposed criteria in the options paper. In light of this, it has been decided that no relocation action will be taken in light of that decision. As indicated in the decisions paper, it has been decided that further work will be done within HMG with regards to potential support – short of relocation to the UK – to those who were included within the database. Whether the Claimants might receive any additional support will be kept under continuing review.”[26]So far as the CX claimants are concerned, the primary decision that was initially under challenge in these closed proceedings was the decision described by Ms Moore in this statement, that no action would be taken to relocate any of the claimants. However, matters have moved on considerably since that decision. It is necessary to take account of subsequent decision-making.[27]On 28 November 2023, Ms Moore made a witness statement in the MP proceedings. She said that on the basis of discussions as to the likely criteria for relocation, it was unlikely that the MoD would take any relocation action in respect of MP1. So far as the MP claimants are concerned, the primary decision that was initially under challenge in these closed proceedings was the failure to take relocation action in respect of MP1, as reflected in the content of Ms Moore’s witness statement. Again, matters have moved on and it is necessary to take account of subsequent decision-making.[28]On 19 December 2023, a policy paper produced by the MoD for the Domestic and Economic Affairs Committee of Ministers (“DEAC”) recommended that a risk-based approach should be taken to establishing eligibility for relocation to the United Kingdom. Case workers had identified approximately 3,000 individuals who had supplied evidence indicating that they were high-profile based on their former employment. [Redacted] officials concluded that additional conditions should be applied in order to prioritise individuals, and those conditions should not be based on risk alone. They recommended that an additional condition should be implemented requiring that, prior to the data leak, the individual “worked alongside, in partnership with, closely supporting or assisting a UK Government department, including where that department can corroborate the link.” It can be seen that this condition closely mirrors part of category 4 of ARAP. Such individuals are likely to have been found ineligible for relocation under ARAP on the grounds that they did not contribute to the UK’s military or national security objectives. There were approximately 200 individuals who were high-profile and who held a confirmed link to the UK Government.[29]The MoD assessed that it would not be feasible to relocate a wider cohort of high-profile individuals because of the “significant operational challenges it would create”. Following consideration of this paper, DEAC agreed that offers of relocation to the United Kingdom should be made to “a targeted cohort of c.200 high-profile individuals and their dependants who hold existing and confirmed links to the UK Government.” It was also agreed that the MoD would conduct further work on a “justifiable means of further prioritisation within the cohort of c.3,000 individuals identified in the Ministry of Defence’s paper and across the entire affected cohort, by the 9th January.”[30]Subsequently, it was assessed that it would be possible to process [redacted] from Afghanistan, and to relocate, from Pakistan to the UK, [redacted]. It is not necessary to set out the detail which leads to this assessment. It is clear, on any view, that only a tiny proportion of those at risk can relocate from Afghanistan each month.[31][In January 2024, an updated threat assessment was carried out by MOD. That covered matters including: - The Taleban would have an interest in the dataset and would take steps which would lead to their obtaining it. - The threat to persons whose details are contained in the dataset is that, if identified by the Taleban, they are at risk, including of lethal harm. - The likelihood that the Taleban will target family members of individuals in the dataset. - The threat would exist regardless of the eligibility for the ARAP scheme of individuals in the dataset. - The risk for ARAP cohorts in third countries.][32][The MOD produced a list containing examples of high-profile roles.][33]On 22 January 2024, the MoD prepared a paper for consideration by DEAC. The paper refers to the threat assessment that all affected individuals, and their families, were at heightened risk as a result of the data leak (it appears that an earlier iteration of the threat assessment was before DEAC, but nothing turns on that). The paper suggests that there were 2,800 individuals who were in high-profile roles (14,000 including dependents, with a further 9,900 individuals if additional family members, who were also at risk, were included), [redacted]. DEAC commissioned further work on a full range of options.[34]On 18 February 2024, the Deputy Prime Minister took the view that insufficient progress had been made to allow a further DEAC meeting to take place. The meeting eventually took place on 25 March 2024 (the day before the hearing). It was supported by a detailed paper from the MoD. That paper made the following recommendations (which were agreed by DEAC):
“a. To offer access to apply for relocation to the UK for the remaining highest risk cohort (an estimated 2,300 individuals, or 11,500 including immediate family members) as a result of the data incident. Such an offer has already been approved for an estimated 168 individuals at highest risk, who also hold existing and confirmed links to HMG. The package will mirror ARAP. If all in this cohort accept the offer, MoD estimates this will extend the current relocation plans [redacted]. b. To adopt a coherent approach to [additional family members (“AFM”)] in light of the data incident, including reassessing AFM applications from the highest risk affected ARAP individuals and introducing a provision allowing individuals eligible under the new route to apply for the relocation of their AFM. Eligibility will be assessed on a case-by-case basis. This is estimated to result in up to approximately 4,950 and 7,400 AFM, respectively. c. That individuals outside of the highest risk cohort (an estimated 13,500 individuals, or 67,500 including immediate family members) will not be offered relocation to the UK. d. To prepare to issue protective security notifications to affected individuals at “break glass”. e. Whether to engage the Taleban at “break glass” to seek to reinforce its amnesty for the affected cohorts. f. To continue to explore the practicalities of making financial payments to affected individuals as a means of reducing the risk introduced by the incident, alongside separately exploring options for a possible compensation scheme. g. To continue to take all possible containment action.”
[35]The paper sets out a detailed analysis to support each of these recommendations. It states that “an estimated 2,500 individuals formerly held high-profile roles placing them at highest risk of Taleban targeting.” After allowing for those who had been approved for relocation under ARAP there were approximately 2,300 further individuals within this cohort (hence the figure in recommendation a).[36]The paper says that officials considered offering an opportunity to apply for relocation to all of those affected by the data leak (almost 100,000 people, including family members). It recognises that, if it were possible to achieve, this would most effectively meet the objective of reducing the risk caused by the data leak. However, it points out that there are constraints. These include the practical difficulties in individuals relocating from Afghanistan, and resettlement challenges in the UK (including the cost and undesirability of using hotel accommodation). It would take [redacted] years to relocate the entire cohort, at a cost of £2.5bn - £3bn. There is an acute housing shortage in the UK, and there are existing pressures on local Government. [Providing a route to the United Kingdom for significant numbers of those named in the dataset may itself give rise to a risk that the fact of the data leak will become known]. It is therefore necessary to prioritise those at the highest risk.[37]As to those who were not in Afghanistan, the paper records:
“Those not in the highest risk cohort who already reside outside of Afghanistan may be easier to assist in practical terms, but the risk of inadvertent disclosure remains. Moreover, that group are not at the same level of risk because they are not in Afghanistan, therefore the most effective action in terms of mitigating risk should be aimed at those at highest risk.”
Grounds of claim Decision under challenge[38]The decisions under challenge in these cases were, initially, the decisions made in November 2023 not to relocate the claimants (or, in the case of MP1, the decision that it was “unlikely” that relocation action would be taken).[39]The substantive and operative decision that is now under challenge is the decision of 25 March 2024 not to include the CX and MP claimants within those who are to be relocated.

Grounds of claim

[40]In each case, the grounds of challenge are as follows:(1) The policy under which the decision was made is irrational.(2) The decision-maker unlawfully fettered their discretion by allowing the policy to automatically determine the outcome of their decision.(3) The decision-maker failed to take into account relevant considerations, namely the personal circumstances of the claimants, when reaching the decision. Submissions[41]In each case, the rationality ground is put in a number of different ways. The evolution of the decision-making means that not all of these are now maintained. The central theme of the remaining complaint is that the targeted cohort who will be offered relocation has been identified by reference to narrowly drawn criteria which preclude an individualised assessment of risk. In particular, no individual assessment has been made of whether MP1 and CX6 fall within the group of those who are at highest risk. Martin Goudie KC, on behalf of the special advocates, says:
“[d]evising a policy which does not allow for consideration of the increased risk that the Claimants face as a result of the data breach combined with their pre-existing relevant personal circumstances is irrational”
. Further, Mr Goudie submits that a significant factor in the decision-making has been the practical difficulties in individuals relocating from Afghanistan. Yet, a significant minority of those named in the dataset have left Afghanistan. Those practical difficulties do not apply to them. Mr Goudie says that the defendants have acted irrationally by excluding those individuals from the targeted cohort without rational justification.[42]Sir James Eadie KC, for the defendants, submits that this is an entirely ex gratia policy which is governed by the prerogative rather than any statutory framework. It is not, therefore, open to the special advocates to advance a challenge on the basis that the decision-makers have fettered their discretion: R (Sandiford) v Secretary of State for Foreign and Commonwealth Affairs [2014] UKSC 44; [2014] 1 WLR 2697 per Lord Carnwath and Lord Mance at [50]. It is open to the special advocates to seek to challenge the decision-making on rationality grounds, but the threshold to show irrationality is high: Sandiford at [66]. The decision-making involves difficult judgments on matters that are for senior Government decision-makers, exercising a very broad discretion, to evaluate, which they have. It is not for the court to second guess their evaluative assessment, or to treat the recently expanded group as a basis for arguing for a yet further expansion of the categories. Discussion The context of the defendants’ decision-making

Discussion

[43]There are important contextual features to the first defendant’s decision-making. Most prominent of these are(1) the existence of a risk to life and a risk of torture, to(2) almost 100,000 people, as a result of(3) a strongly arguable breach of the law on the part of the Government, in circumstances where(4) those affected are not aware of what has happened and are disabled from taking action to protect themselves, because(5) the Government has, highly exceptionally, secured a super injunction on the basis of a promise to the court that it would take all practical steps to protect those at risk as soon as possible, in circumstances where(6) the decision-making is fraught with risk and involves the allocation of significant resource, potentially to the detriment of others, and(7) there is no statutory framework regulating the decision-making.[44]As to (1) and (2), the decision-making that is challenged by the special advocates is a response to the data leak. That leak was of the utmost seriousness. The first defendant’s assessment is that it gives risk to a risk of torture and a risk to life of almost 100,000 people (including family members) who have sought the assistance of the UK Government.[45]As to (3), although the data leak was accidental, on the materials before us it is strongly arguable that it also involved a breach of the law. That is because the data comprised personal data within the meaning of sections 3(2), (3) and 14(c) of the Data Protection Act 2018, and article 4(1) of the UK General Data Protection Regulation. The first defendant is the data controller (within the meaning of the Act and Regulation) because he holds the authority to determine the purposes and means of processing the personal data. [Individual 1] was a data processor (within the meaning of the Act and Regulation) because [they were] provided with the dataset and [they] used it and disseminated it. There was an obligation on the first defendant, [Individual 1], and all MoD officials to ensure that appropriate protection was in place (including by means of appropriate technical or organisational measures) against unauthorised or unlawful processing of the dataset, and against accidental loss: article 5(1)(f) of the Regulation (and see also articles 24(1), 25 and 32). It required no hindsight to appreciate that unauthorised disclosure of the dataset might give rise to a risk to life. It should never have been provided to, or disclosed by, [Individual 1], in unencrypted form. The provision of the dataset to, and its disclosure by, [Individual 1], involved a clear breach of article 5(1)(f) (and also article 24(1), 25 and 32) of the Regulation.[46]As to (4), the first defendant has taken positive steps to prevent the claimants from becoming aware of what has happened. The claimants are thus disabled from taking the action to protect themselves that they might otherwise take. In this sense, and to this extent, the first defendant’s actions have not only caused the risk to arise, but have also then materially increased one aspect of the risk (recognising that the first defendant’s assessment, which is not challenged by the special advocates, is that the balance of risk justifies the non-disclosure of the data leak to the claimants). In circumstances where the claimants are not (for what it is common ground at the moment is a good reason) being informed of the data leak, the court must scrutinise the steps taken by the first defendant to protect the claimants.[47]As to (5), the injunction that has been granted, and maintained, is truly exceptional: [2023] EWHC 2999 (KB) per Chamberlain J at [2]. It was sought and granted in private proceedings; it is made against the entire world; it has not been disclosed to those who have the greatest interest in it (namely the claimants); its mere existence is, currently, a secret and may not be disclosed without incurring potential liability to imprisonment; it has been maintained, in this form, for 7 months. It is, so far as we are aware, the only order of its type that is in existence. It is without any clear precedent that we are aware of, since the limited jurisdiction to make non-disclosure orders was clarified in Practice Guidance issued by the Master of the Rolls in 2011 (see per Chamberlain J at [24]). The application for the injunction was made for the purpose of enabling the first defendant to take (over a period that was initially assessed to be just 4 months) “what steps he can to protect those named in the data set”. The witness statement in support of the application, made by Ms Cope, the first defendant’s chief operating officer, stated that the application was sought to enable the first defendant to undertake “urgent targeted protective action to mitigate the risk to life and risk of persecution to the people named in the breach.”[48]As to (6), any positive action by the first defendant has to take account of the potential that this would increase the risk to the claimants. Most notably, it must take account of the potential that positive action could result in the Taleban becoming aware of the data leak (with the assessed consequence that they would obtain and exploit the dataset, directly giving rise to a risk to the lives of almost 100,000 people). The current assessment is that informing those named in the dataset of the leak would materially increase the risk to them. In the absence of cross-examination, we are not in a position to test that assessment. The super injunction has been granted because of this risk. For the same reason it may not currently be practicable (at least not without very careful work) to put the claimants in a position where they could take steps to protect themselves, for example by informing them of the risk and by providing funds to enable them to take their own protective steps. Providing a route to the United Kingdom for significant numbers of those named in the dataset may itself give rise to a risk that the fact of the data leak will become known. Further, the decision-making involves a series of evaluative assessments that require consideration of wide-ranging policy factors. They concern questions of international relations, macro-economic and social policy, immigration policy (including who should be permitted to enter the United Kingdom) and the balancing of competing interests and limited resources. For example, providing a route to the United Kingdom to a large number of people would give rise to consequential obligations to provide accommodation and support for those who arrive. That is likely to involve a significant allocation of limited resources, potentially to the detriment of others. These are all matters in respect of which the first defendant is the primary decision maker. We accept the submission of Sir James Eadie that they have a very broad discretion when setting and applying the applicable criteria.[49]As to (7), it is common ground that the first defendant has no statutory duty to mitigate the damage that has been caused by the data leak. Any policy response to the data leak is, in that narrow sense, ex gratia. The formulation of a policy response is governed by the prerogative. It is common ground that the defendants’ decision-making can be reviewed on grounds of rationality: R (Abbasi) v Secretary of State for Foreign and Commonwealth Affairs [2002] EWCA Civ 1598; 2003] UKHRR 75 per Lord Phillips MR at [106], Sandiford per Lord Carnwath and Lord Mance at [50].[50]Where a policy is adopted under an applicable statutory framework, it may be unlawful for the policy to be framed, or to be operated, in a blanket manner that permits of no exception to the strict terms of the policy (unless, possibly, the decision maker is willing to consider changing the policy in response to new situations). That is because such an approach is likely unlawfully to fetter a statutory discretion: British Oxygen Co Ltd v Board of Trade [1971] AC 610 per Lord Reid at 625D, R v Secretary of State for the Home Department ex parte Venables [1998] AC 407 per Lord Browne-Wilkinson at 496-497. Where, as here, there is no governing statutory framework, and the applicable policy is adopted under prerogative powers, there is no applicable statutory discretion and thus there is no question of an unlawful fettering of a discretion in that sense: Sandiford at [62]. However, it remains the case that decision-making is reviewable on grounds of rationality. That can, in principle, encompass the question of whether it is rational to adopt a rigid rule which precludes the consideration of the position of individuals on a case by case basis.[51]On the other hand, the special advocates point out that the only reason why the decisions are being made under prerogative powers is because the background cannot be made public. Were it otherwise, it is likely that (as is the case with ARAP) the policy would be implemented by way of changes to the Immigration Rules. Such changes would be introduced under section 3(2) Immigration Act 1971, would be required to be laid before Parliament, would be subject to the negative resolution procedure, and would be subject to the constraints imposed by the 1971 Act.[52]As to the application of a rationality test in a prerogative context, both parties relied on the decision in Sandiford. In that case the appellant was facing the death penalty in Indonesia. She challenged the Government’s blanket policy of refusing to provide funding for legal representation for British nationals facing criminal proceedings abroad. The court held that because the Secretary of State’s power to provide assistance derived from the prerogative, it was not subject to any statutory constraint and it was therefore permissible to adopt a rule admitting of no exception, subject to the rule satisfying a test of rationality. Both parties relied, in particular, on the observations of Lord Carnwath and Lord Mance at [66]:
“‘Irrationality’ is a high threshold, but it may be easier than otherwise to surmount in a case involving an imminent risk of death by execution of a British citizen deprived of financial support abroad. The court’s role is given added weight in a context where the right to life is at stake (see R (Bugdaycay) v Secretary of State for the Home Department [1987] AC 514). A keen scrutiny of the policy and its application must on any view be required in such circumstances. There may be scope in an appropriate case to test the legitimacy of the blanket policy that the Foreign Office currently advances, by reference to a broader framework of proportionality discussed in a non-Convention context in Kennedy v Information Commissioner [2014] UKSC 20; [2015] AC 455.”
[53]This case, like Sandiford, involves a risk of death, but it also has the other contextual features that we have identified. Sir James Eadie, with some understatement, accepted that these factors resulted in the case being “some way up the hierarchy of cases where ‘anxious scrutiny’ is required by the court.” He accepted that, in this context, anxious scrutiny involves the court closely analysing the decision to determine whether it provides a sufficiently cogent and reasoned justification for the action taken, that withstands close and careful logical scrutiny: see R (Hoareau) v Secretary of State for Foreign and Commonwealth Affairs [2020] EWCA Civ 1010; [2021] 1 WLR 472 at [153] – [156].[54]That is the approach that we take. It is entirely consistent with our recognition that the first defendant is the primary decision maker and is entitled to a wide measure of evaluative discretionary judgement, but it also gives appropriate weight to the other contextual factors that we have identified.

Rationality challenge

[55]The primary ground of challenge in each case is that the policy that has been adopted is irrational. Such a challenge may succeed where there is a demonstrable flaw in the reasoning: R (Law Society) v Lord Chancellor [2018] EWHC 2094 (Admin); [2019] 1 WLR 1649 per Leggatt LJ and Carr J at [98]. The intensity of review to be applied in order to determine whether there is such a flaw depends on the context. For the reasons we have given, the contextual features in this case requires a particularly close analysis of whether the decision-making provides a sufficiently cogent and reasoned justification for the action that has (and has not) been taken.[56]We are content to assume that the first defendant is justified in concluding that it is not practicable to relocate all those at risk to the United Kingdom. [Redacted] The risks that such an exercise would be designed to prevent are likely to materialise well before the exercise could be completed.[57]It necessarily follows that very difficult decisions have to be made as to who should and who should not be relocated. [Redacted] the first defendant is entitled to conclude that high-profile individuals in the roles that they have identified fall into the highest risk category, and that they should therefore be prioritised for protection (subject to making appropriate alternative provision for the rest). [MOD does not exclude the possibility that high-profile individuals occupying other roles may also fall into the same highest risk category. That is because the highest risk category is concerned with those occupying high-profile roles and MOD gives only a non-exhaustive list of examples of such roles.] [Redacted] The strong implication is that other high-profile people holding other roles will be at equivalent risk, depending on their particular circumstances. This is more than mere semantics. Given the very particular circumstances of these proceedings, and the absence of any input from the claimants, this court has to apply anxious scrutiny to the way in which [decisions have been taken].[58]Yet it appears that the decision-making has proceeded on the basis that only those who claim to have held one of the specified roles can qualify for relocation. There is no evidence that it would be impracticable to give individual consideration in each case to determine whether another high-profile person, albeit with a different role, falls within the highest risk category. Each person in the dataset has made an ARAP application. The first defendant therefore has at least some information about each applicant on which to assess whether they are high-profile, and whether their role will expose them to the same risk as those in the identified roles. We leave open the question as to whether it is necessary to give separate and individual consideration to every ARAP applicant because it is at least conceivable that there will be some persons who could not on any view be considered “high-profile.” That is not the case for either the judges or journalists who are the subject of these proceedings.[59]The position can be tested by reference to MP1. He is a former Afghan judge. His work as a judge is explained in detail by Julian Knowles J in his judgment at [86] – [95]. He was the chief judge of the Primary court in Helmand Province. He then became an appeal court judge. He was the only permanent judge in Helmand Province. He dealt with crimes including murder and violence against women. The perpetrators were often members of the Taliban. He sentenced and imprisoned Taliban members who subsequently obtained high positions in the present regime. In 2020, there was an attempt on MP1’s life, for which the Taliban took responsibility.[60]The first defendant has (implicitly) concluded that MP1 did not hold a high-profile role. He has (implicitly) done so because a “judge” is not one of the roles listed [Redacted]. [That, however, wrongly assumes that it is an exhaustive list of those roles that should be treated as high-profile for these circumstances.]. The first defendant has failed to give any explicit consideration to whether MP1’s role was high-profile. It is not for us to determine whether MP1’s role was high-profile: that is for the first defendant. But he must do so on a basis that withstands logical scrutiny. So far, he has not done that.[61]Much the same applies to CX6. He was a journalist. He worked for the BBC World Service and was the presenter of a well-known popular radio programme which gave panellists and audience members a platform to discuss issues of national importance. It can readily be seen that CX6 may well be thought to have had a high-profile. Whether he held a high-profile role [redacted] is for the first defendant to determine. Again, so far, he has not done that.[62]For these reasons alone, the challenge in respect of MP1 and CX6 succeeds. We grant permission for them to seek judicial review, and we quash the first defendant’s decision not to relocate them. We will make an order requiring the first defendant to reconsider their cases.[63]There is no need separately to consider the position of claimants MP3, MP4 and MP5. They are not separately named within the dataset. They are family members of MP1 and their position is contingent on the decision in respect of MP1. If the first defendant lawfully decides not to offer to relocate MP1, it will inevitably follow that MP3, MP4 and MP5 will not be relocated. If the first defendant offers to relocate MP1 then it will be necessary for the first defendant to consider whether to extend that offer to MP3, MP4 and MP5.[64]Nor is there any need separately to consider CX1 and CX2. The latter has withdrawn his ARAP application. The former has accepted relocation to Canada and it can be taken that his ARAP application has also been withdrawn.[65]That leaves CX4 and CX7. They are not named in the dataset. The data leak does not therefore, directly, give rise to any risk to them. Nor does it, directly, give rise to a need to consider steps to mitigate risk. The special advocates suggest that they might be at risk because of a possible professional link to CX6. There is, however, no evidence of such a risk. It is altogether too speculative to require the first defendant to take further action in respect of them. Were it otherwise, the first defendant would be required to investigate risks not just to those named on the dataset and their immediate and extended families (as they currently are doing) but also any individual who has any link to those named on the dataset. That would be an enormous and unmanageable undertaking. The fact that the first defendant has not embarked on such an assessment cannot, on the evidence, be categorised as irrational. Other matters[66]The special advocates raised other challenges to the decision-making, beyond the stark and central rationality challenge. In the light of the outcome of that challenge it is not necessary for us to address those other challenges. To a large extent, they have been overtaken by events, because they have been superseded by the first defendant’s decision of 25 March 2024 which was (by the time of the hearing on 26 March) rightly the focus of the special advocates’ challenge.[67]There is, however, one aspect of the decision-making that has caused us particular concern and which it is appropriate to mention. The focus on identifying the highest risk cohort may have meant that insufficient attention has been given to the steps that should be taken to protect the rest. They also face significant risks. The potential risk to them is of the same kind as the potential risk to those who are being prioritised: it is a risk of torture and/or death. [Redacted]. [The distinction that suffices for the purposes of identifying the highest priority cohort would be an insufficient basis for disregarding the rest]. In fairness, the first defendant has not disregarded the rest. The latest decision-making shows that considerable work has been undertaken to assess what can and should be done in respect of all of those on the dataset (and their families) who will not be offered a route to the United Kingdom. It is, however, now 7 months after the application for an injunction was made, when a timeframe of 4 months was given for remedial action. Even now, it is far from clear precisely what will be done in respect of those who are not prioritised for relocation. It may well not be possible to do very much in advance of “break glass”, because of the risks that positive actions will themselves precipitate “break glass”. But there should, at least, be a plan in place that can be implemented as soon as “break glass” occurs.[68]The first defendant has undertaken work on the communications that it might be appropriate to undertake following “break glass”. However, it does not appear that much work has been undertaken to explore the provision of financial resources to enable individuals to take their own measures to protect themselves. The MoD paper of 25 March seeks the agreement of the DEAC “to continue to explore the practicalities of making financial payments to affected individuals as a means of reducing the risk introduced by the incident”. That does not seem to us to recognise, sufficiently, the urgency of having a package of measures in place ready to be implemented as soon as “break glass” occurs.[69]Further, there does not seem to us to have been a sufficient focus on those who are not in Afghanistan. Some of those will be in sufficiently safe third countries with no risk of refoulement to Afghanistan. But others may be unsafe. Again, MP1 is an example. He lives in Pakistan, but close to the border with Afghanistan. There are two particular risks that do not seem to us to have been fully taken into account. One is that he may be subject to refoulement to Afghanistan. If so, he will then immediately and directly come into contact with the Taleban authorities. [Redacted].[70][Further, an important reason for prioritising those at highest risk is that it is not possible for all those at risk in Afghanistan to relocate. That issue does not apply to those, like MP1, who are not in Afghanistan]. It seems to us, therefore, that the first defendant has not, so far, given sufficient attention to the steps that can and should be taken to protect those who are in [redacted] third countries and who may be at risk of refoulement. Outcome[71]So far as MP1 and CX6 are concerned, we grant permission to claim judicial review of the decision of 25 March 2024 on the ground that the decision is irrational. We quash paragraph 2c of the decision of 25 March 2024 not to offer relocation to the United Kingdom to individuals outside (what the first defendant has determined to be) the highest risk cohort. A rule which categorically excludes anyone not in the identified roles is irrational in its application to the claimants.[72]We direct that the first defendant should reconsider the approach to be taken to identifying those who are within the highest risk cohort.

Cited in 4 later judgments