Ministry of Defence v Global Media and Entertainment Limited & Ors

On 1 September 2023, a super-injunction was granted contra mundum (against the world) on the application of the Ministry of Defence (“MOD”). The super-injunction prevents disclosure of:

The original decision to grant the super-injunction was made by Robin Knowles J on the basis that the acquisition of the dataset by the Taliban would expose a very large number of people, mainly in Afghanistan but also in other countries, to the risk of severe ill-treatment and/or death.

The subsequent decisions to continue the super-injunction in force, with modifications, on 2 November 2023, 1 December 2023, 18 December 2023 and 15 February 2024, were all mine. The super-injunction continues in force at the date of this judgment. It has therefore been in force for more than 8 months. There is no reported example of a super-injunction continuing in force for so long. There is also no reported case whose facts are remotely comparable to those here, or where the potential harm the injunction seeks to prevent is as grave.

The reasons why the MOD first sought an injunction are set out in my private judgment of 23 November 2023 (neutral citation [2023] EWHC 2999 (KB) (“Judgment no. 1”)), at [5]-[8], and are not repeated here. The procedure adopted has been complex because no-one who knew about the injunction initially opposed the relief sought and because of the need to deploy sensitive material (i.e. material whose disclosure would cause damage to national security), in order to ensure that the Court was properly informed.

At first, the law officers appointed an Advocate to the Court to make submissions on that material: see Judgment no. 1, [12]-[17]. Then, the media organisations, which had at first been content to observe but not participate in the proceedings, applied to be joined. I granted their applications and they are now defendants to the claim. A declaration under s. 6 of the Justice and Security Act 2013 was then made and the Advocates to the Court became Special Advocates: see my private judgment of 15 February 2024 (neutral citation [2024] EWHC 312 (KB) (“Judgment no. 2”)), at [2]-[10].

I heard submissions at each stage from some of the journalists who know about the subject matter of these proceedings: see Judgment no. 2, [2]. However, at the hearing on 1 February 2024, it became clear to me that submissions were not enough. There were aspects of the MOD’s written evidence which required further investigation. Accordingly, I gave directions for:

The evidential hearing was due to take place on 27 and 28 March 2024, but key evidence was filed late and key recommendations relevant to the facts were made by the Domestic and Economic Affairs Committee of the Cabinet (“DEA”), which is chaired by the Deputy Prime Minister and includes a number of senior Cabinet Ministers, very shortly before the hearing. The parties and the Special Advocates agreed that a new date would have to be set.

The media organisations have now filed their own evidence in the form of a witness statement from Pia Sarma, the Editorial Legal Director of Times Media Ltd., which contains information relayed to her by journalists from all the media organisations. There was a hearing on 30 April and 1 May 2024 at which Natalie Moore, who heads the Defence Afghan Relocation and Resettlement Directorate (“DARR”) at the MOD, was questioned by Mr Bunting KC (for the media organisations) in private and by Tom Forster KC (lead Special Advocate) in CLOSED. The MOD were represented by Sir James Eadie KC and Cathryn McGahey KC. I am grateful to all counsel for their questions and submissions, to the journalists and media organisations for acting in a wholly responsible way throughout these proceedings and to Mrs Moore, who gave frank and helpful answers to the best of her ability, reflecting the MOD’s corporate knowledge.

This judgment records the reasons for my decision, in the light of the evidential hearing on 30 April and 1 May 2024, that the injunction should now be discharged with effect from 21 days after the handing down of this judgment.

I set out the legal principles in Judgment no. 1 at [24]-[39]. All parties now accept that these are the principles to be applied.

There has never been any dispute that the fact of the data breach is information whose disclosure would give rise to a real risk to life and damage the UK’s national security; and that this information is in principle protectable by an injunction: see Attorney General v BBC [2022] EWHC 826, [2022] 4 WLR 74, [19]. It is common ground that it is not enough for the MOD to show that disclosure of the information protected by the injunction would give rise to a real risk to life. The Court must also consider whether the grant of relief might endanger life, whether directly or indirectly and, if so, to balance these consequences: see Judgment no. 1, [35].

It was also agreed on all sides that, although due weight must be given to risk assessments made by the executive, it is for the Court to conduct this balancing exercise. In this respect, these proceedings are different from judicial review proceedings: see Judgment no. 1, [30]-[31], citingmy judgment in Attorney General v BBC, [30]-[33].

At this stage, I should say a little more about what “due weight” means in this context. In most cases, the decision whether to grant interim relief will be taken on the basis of written evidence alone. That has been the position in these proceedings at all stages prior to the last hearing on 30 April and 1 May 2024. However, given the gravity of the potential consequences of my decision and the numbers affected, I was not satisfied that the written evidence before me provided a sufficiently robust basis for conducting the balancing exercise required. That is why I gave the directions I have described, leading to the evidential hearing: see Judgment no. 2, at [33]-[36].

The result is that the evidence before me is fuller than at any stage previously. I have the benefit of Mrs Moore’s oral evidence. The evidence still includes assessments of risk made with the benefit of institutional expertise, to which I must attach weight, but I am now better able than I was to judge how much weight should be attached to assessments on particular issues.

It was a prominent theme of Mr Bunting’s submissions that both the justification for the injunction and the predictions about its potential duration had developed materially over time. In my judgment, he was correct to draw attention to this aspect of the case.

When the application for an injunction was first made, Nina Cope said this in her witness statement of 29 August 2023, at para. 28:

“The Claimant is making his present application for an injunction as it is believed that this is the best and most effective way to gain sufficient time to implement protective measures to reduce the risk to the lives of those affected by the unauthorised data release.”

It was anticipated that it would take “in the region of four months” for “all reasonable mitigations” to be put in place (para. 41). That explains why Robin Knowles J set a return date of 1 December 2023.

However, the prospect of a super-injunction being in force for as long as 3 months led Nicklin J, then the Judge in Charge of the Media and Communications List, to direct an earlier hearing, in October 2023, which was in due course listed before me.

At each of the hearings before me, the application for relief has been advanced on the basis that extra time was required to undertake further work to mitigate the effect of the data breach; and the expectation has been that the justification for the injunction would likely fall away in a matter of weeks, or at most a few months, because the Taliban were likely to find out about the data breach and then gain access to the dataset. This eventuality was termed a “break glass” event. The MOD assessment that the Taliban have not learned about the existence of the data breach, and have not obtained access to the dataset, has remained unchanged throughout. I shall return to this point later. The important point for present purposes, however, is that, on the MOD’s current case, there has not been a “break glass” event for some 8 months, and it must follow that the injunction may well be required for a very substantial further period.

Meanwhile, decision-making on how to protect those whose details were included in the dataset (“Affected Persons”) who are not eligible under the Afghan Relocation and Assistance Policy (“ARAP”), and their family members, has taken place with no opportunity for public, media or Parliamentary scrutiny and at what the media defendants submit is a slow pace.

By 10 October 2023 (nearly half-way through the period between 1 September and the original return date of 1 December 2023), no decisions had been taken to address the position of Affected Persons who were ineligible under ARAP (see Deana Rouse’s witness statement of 10 October 2023, para. 18). This remained the position on 23 October 2023 (see Deana Rouse’s witness statement of 23 October 2023, para. 15).

At a meeting on 16 November 2023, the DEA directed officials to prepare plans to relocate a targeted cohort of c.150 ineligible Affected Persons and their family members, but stipulated that officials “must revert to the committee if there is a significant increase in the size of this targeted cohort” (witness statement of Natalie Moore of 28 November 2023, para. 10).

The DEA met again on 19 December 2023 and agreed that relocation should be offered to “a targeted cohort of c.200 high profile individuals and their dependents”. The “Actions and Decisions” document records the DEA’s view that “hotels would not be required to accommodate this cohort”. As I noted in Judgment no. 2, at [12(b)], one inference that might be drawn from this is that the need to avoid the use of hotels (which the Government has publicly committed to phasing out) played a part in the decision about the size of the cohort to whom relocation should be offered. Mrs Moore accepted in evidence that Ministers had at that stage given a strong steer that hotels should not be used. Although there is no absolute prohibition on the use of hotels, this remains the Government’s position.

At the DEA’s meeting on 22 January 2024, a decision was taken to “consider how we could deliver support” to a cohort of 2,800 high profile individuals, although at that time relocation was considered impractical: see Judgment no. 2, [12(d)]. Since then, there has been a further DEA meeting on 25 March 2024, at which it was agreed to offer access to apply for relocation to the UK to the remaining highest risk cohort, an estimated 2,300 individuals or 11,500 including immediate family members, and to adopt a “coherent risk-based approach to Additional Family Members”.

The papers before the DEA set out, among other things, the costs of relocating this cohort (“the relocation cohort”). Taking into account the costs of housing them in the UK, the costs will be spread over years. The precise amount of these costs depends on a series of policy decisions not yet taken. For present purposes, it can be assumed that they will run into several billion pounds: the sort of money which makes a material difference to Government spending plans and is normally the stuff of political debate. There is a real question about the feasibility (let alone the desirability) of keeping the reasons for such expenditure secret, given the established processes for scrutiny of public expenditure. In any event, as noted above, the effect of the injunction is that, to date, the decision to incur these costs has been subject to no public, media or Parliamentary scrutiny.

The DEA also endorsed a decision that “individuals outside of the highest risk cohort will not be offered relocation to the UK”. This last decision is now being reconsidered: see Natalie Moore’s witness statement of 23 April 2024, para. 6. But there is no suggestion that this reconsideration will lead to a substantial increase in the cohort to whom relocation will be offered. The evidence establishes that there are about 80,000 “individuals outside the highest risk cohort” (including immediate and additional family members) for whom the Government proposes (at this stage) to do nothing, apart from continuing to develop a communications strategy in the event of a “break glass” event. Because of the injunction, there has been no opportunity for argument about how to balance the costs of providing other assistance to this cohort against the moral obligation owed to those whose lives have been put at risk by conduct for which the UK Government is responsible.

To date, while the evidential picture has been evolving, I have taken the decisions to extend interim relief one by one, asking on each occasion whether a limited extension is justified for the time being. I considered that approach to be justified while the evidential picture was evolving. However, an incremental approach carries with it the danger that one loses sight of the overall duration, and likely further duration, of the relief granted. Now that the evidential picture has become relatively stable, it is important to take a realistic view of the logical end-point of the process if the injunction is maintained – and to ask whether the maintenance of interim relief for such an unprecedented period can be justified.

By the time of the hearing on 30 April and 1 May 2024, the following was common ground:

The nettle must therefore be grasped now.

So far as the injunction itself was concerned (i.e. whether the fact of the data breach should continue to be protected from disclosure), the key factors on which I based my decision were set out in Judgment no. 1 at [43]:

“…the risk assessments… are appropriately caveated. There are a number of imponderables. However, they nonetheless contain assessments which indicate a real risk that: (i) the Taliban do not already know about the compromise of the dataset; (ii) disclosure of the fact of its compromise would cause them to take steps which lead to their obtaining it; and (iii) in that case, many thousands whose details are included in the dataset could be killed or injured and the UK Government would have no realistic way of safeguarding them...”

At that stage, no-one was disputing that the injunction protecting the underlying information should be continued, though Mr Forster (then as Advocate to the Court) did submit that the “super-element” (i.e. the element protecting the fact of the injunction’s existence) could and should be discharged.

By February 2024, when I gave Judgment no. 2, the DEA had agreed to offer relocation to the UK to around 200 high profile individuals and their dependants: Judgment no. 2, [12(b)]. There was a real possibility that the injunction had enabled hundreds of ARAP eligible persons to leave Afghanistan since 1 September 2023 in circumstances when they would otherwise have been targeted by the Taliban for extra-judicial killing or serious physical harm. There was a real possibility that the continuation of the injunction might save the lives of 4,000 individuals then in Afghanistan who had been identified for relocation but had not yet been relocated. There was also a further group comprising approximately 2,000 Triples and others whose ARAP eligibility was being reconsidered, whose lives may be saved by the continuation of the injunction: Judgment no. 2, [31(b)].

Against this, there was a much larger cohort who – on the basis of then current Government policy – would never be offered relocation to the UK (“the non-relocation cohort”). Continuation of the injunction might delay the date at which the Taliban would target those in this cohort, but on the MOD’s case that date was very likely to come “in a matter of months” in any event; and there was a risk that continuation of the injunction might positively endanger these people by preventing them from being told about the risk they face and/or by delaying the point at which the MOD can compensate them for the data loss, leaving them unable to take action to avoid the risk: Judgment no. 2, [31(c)]. In addition, there was a risk that the injunction was indirectly harming this large cohort by shutting off the usual mechanisms of democratic accountability. In particular, media pressure might change the decisions taken about what assistance to give to those in this cohort: Judgment no. 2, [31(d)-(f)].

I considered carefully at each stage whether the “super-element” of the injunction could be discharged. The MOD did not initially seek a super-injunction. Nonetheless, I gave reasons, on the basis of the evidence in November 2023, why discharging the “super-element” would risk unravelling the protection afforded by the injunction: Judgment no. 1, [44]-[45]. The position had not changed by February 2024: Judgment no. 2, [37].

The main cohort that the injunction may be serving to protect is the cohort currently in Afghanistan who have been, or may be, offered relocation to the UK. The same points apply to some of those in third countries (notably Pakistan) who are being offered relocation. When immediate and additional family members are taken into account, this group may amount to about 20,000 individuals.

As I said in Judgment no. 1 at [43] (see [29] above), the assessment that the injunction is offering a protective benefit to this group depends critically on two assessments: (i) the Taliban do not already know about the compromise of the dataset; (ii) disclosure of the fact of its compromise would cause them to take steps which lead to their obtaining it.

If the Taliban do not know of the existence of the dataset, it is likely that the injunction is, overall, having a protective effect, for this cohort, because it creates a window during which some of them may be able to leave Afghanistan (or the third countries where they are) before the Taliban obtain the dataset. In assessing the extent of the benefit, however, I bear in mind that the cohort of 2,300 individuals identified in the DEA’s decision document of 25 March 2024 are those in the highest profile groups. Many of these may be well known to the Taliban in any event, even without the dataset, though in some cases the information contained in it may make it easier to locate them.

However, as Mr Goodall (of Global) pointed out through Mr Bunting, if the Taliban already know of the compromise of the dataset then either:

In scenario (a), the injunction is likely to be significantly increasing the risk faced by the individuals in this cohort, because the Taliban have the dataset, but some of the individuals they are targeting do not know this and those of them who might otherwise be able to take action to protect themselves (such as travelling to another part of the country or taking other steps to hide themselves and their families) cannot take that action. In scenario (b), the injunction is probably neutral vis-à-vis this cohort.

All this makes it very important to probe in detail the assessment that the Taliban do not know of the existence of the dataset. I have considered carefully what is said in CLOSED about the basis for that assessment. However, there is a significant possibility that the Taliban already know of the existence of the dataset. I base that view on both OPEN and CLOSED evidence. The OPEN evidence includes the following, in particular:

The latter two points, which emerge from the evidence filed on behalf of the media organisations, seem to me to be of particular significance.

If the Taliban do know of the existence of the dataset, the MOD’s own evidence suggests that scenario (a) is more likely than scenario (b). I have considered the MOD’s view that there is very little that those notified can do to protect themselves. I do not accept that this would be true for everyone. Much would depend on the individual’s financial resources and contacts. The one thing that can be said with confidence is that Affected Persons would be better off learning of the data breach by notification from the UK Government than from a knock on the door by the Taliban. This means that there is, at minimum, a significant possibility that the injunction is actually harming, rather than protecting, the cohort who are in Afghanistan who have been offered relocation and who have not yet begun relocating. The same is true of some of those in other countries. In my judgment, this possibility must be taken into account when assessing the balance of risks.

This is so even if, as the MOD assess, it is on balance unlikely that the Taliban know of the existence of the dataset. Harms which may be caused by an injunction must be taken into account. They cannot simply be excluded from the calculus because they will only arise in a scenario regarded as unlikely on balance. (Otherwise, it would be impermissible to pray in aid in support of the maintenance of the super-injunction the 25-35% chance that discharging the “super-element” would cause harm, as the MOD have consistently sought to do.)

I have considered whether the injunction might also be having another adverse effect on this cohort, by denying them and others on their behalf the opportunity, through the media and Parliament, to put pressure on the Government to increase the rate of relocations (which the MOD describes as “necessarily very limited”: see the MOD’s OPEN skeleton argument, para. 8). The MOD assess that there may be limited steps Affected Persons could take to protect themselves. I say more about this in my CLOSED judgment. However, as I said in Judgment no. 2, at [31(f)], “remedial action honestly considered infeasible at a particular point in time may come to be regarded as possible days or weeks later if, in the intervening period, a media campaign…has caught the public’s imagination.” In my judgment, there is a significant possibility that media and/or Parliamentary pressure might cause options which have to date not been considered feasible to be viewed differently.

The main reason why I decided that the evidential hearing was necessary was to examine further the position of those Affected Persons who will not be relocated. The exact number is unclear, but it is likely to be around 80,000 when immediate and additional family members are taken into account.

Sir James Eadie’s submission was that, even for this group, the injunction is having a protective effect for the time being. On one level, that is right, at least on the assumption that the Taliban do not already know of the data breach: at any given point in time, the members of this group are better off if the Taliban do not have the dataset than if they do. However, the calculus must also take into account the following types of harm, which the continuation of the injunction is liable to cause to members of this group:

The purpose of the super-injunction has been to protect lives. In deciding whether to maintain it, the most important consideration is the balance of risk to the Affected Persons – i.e. whether the protective effect is outweighed by the risks which are or may be caused by the continuation of the injunction. In assessing the latter risks, I have taken into account the possibility that public debate might itself result in measures which enure to the benefit of Affected Persons (or at least the large cohort not currently being offered relocation). However, even in a context where lives are at stake, the restriction on public debate is a matter which deserves some weight in its own right, for the reasons set out in Judgment no. 1, at [35(d)]-[38].

In Judgment no. 1, at [39], I said: “it is axiomatic in our system that decisions subject to public and Parliamentary scrutiny are not only more legitimate, but also likely to be better than ones taken in secret”. The size and import of the decisions now being taken, both in moral and in financial terms, bring this point into sharp relief. It is fundamentally objectionable for decisions that affect the lives and safety of thousands of human beings, and involve the commitment of billions of pounds of public money, to be taken in circumstances where they are completely insulated from public debate.

The MOD’s case for the maintenance of the injunction depends on a series of assessments: some about existing fact (e.g. whether the Taliban already have the dataset) and some in the nature of predictions about what may be done in the future in particular factual scenarios. On the evidence I have seen, OPEN and CLOSED, each assessment has been made in good faith. It is not uncommon for government policy to depend on assessments of this sort. But the use of these assessments in support of the maintenance of the super-injunction is, nonetheless, very unusual in at least three respects.

First, the risk assessments are caveated and there are a number of imponderables. It is tempting to take each assessment as a building block upon which to make further assessments and use this process to generate an overall view as to what is likely to happen if the super-injunction is discharged. But there is a real danger that, in doing so, one builds an edifice with very unsure foundations, ignoring the consequences that may be ensuing or may yet ensue if some of the initial assessments are wrong.

Second, even on the MOD’s assessments, there is a significant risk that the relief granted to date has in fact put lives at risk and is still doing so. This risk must be factored into the balance of risks and benefits.

Third, the relief granted to date not only prevents public discussion of the full reasons for the Government’s policy. It prevents the public from knowing of the very existence of the policy or the problem which it addresses; and it deprives decision-makers of information, public and Parliamentary scrutiny, all of which are liable to improve the quality of their decisions.

At this stage, it is important to return to the principles underlying the grant of interim non-disclosure orders: see Judgment no. 1, [24]. Interim non-disclosure orders of any kind require a “clear and cogent” evidential basis. That applies with even more force where the order is a super-injunction, which must be justified on the basis of strict necessity; and the maintenance of a super-injunction for any longer than a short period is “truly exceptional”.

In my judgment, the MOD has not shown that the maintenance of the injunction is now justified, given where the balance of risks now falls and the effect of the injunction in closing off public debate on an issue of profound moral and economic significance. In particular:

For these reasons, I propose to discharge the injunction with effect from 21 days after the handing down of this judgment. That period is designed to allow for protective notifications to be sent to Affected Persons in Afghanistan and third countries and to allow time for an appeal. If the MOD decide to apply for permission to appeal, it may be that a further extension of the injunction is considered appropriate, but the Court of Appeal (which will know better than I do when it can hear any appeal) is best placed to decide on any such extension.

In the light of my decision, the question whether to discharge only the “super-element” of the injunction does not arise. However, in case the point becomes relevant on appeal, I record that, had I been disposed to continue the injunction at all, I would have discharged the “super-element”. Even though the discharge of the super-element would not enable public scrutiny of the underlying decision-making, it would enable public debate about the use by the courts of super-injunctions in proceedings brought by the Government. We have now reached a point where the continuation of the super-injunction for more than eight months represents a wholly novel use of the remedy, which is itself a matter of significant public interest. The assessed risk (falling short of a likelihood) that the discharge of the super-element would unravel the protection afforded to the fact of the data breach is no longer sufficient to justify the continued stifling of public debate on this issue. The fact that discharge of the super-element alone would leave the debate less than fully informed is not a sufficient reason to maintain in force a legal prohibition whose effect is to stifle it completely.

The draft of this judgment was sent for security checking pursuant to CPR 82.17 on 14 May 2024 and was amended pursuant to representations made under that rule.

My intention is to publish this judgment, and Judgment nos 1 and 2, on the day on which the injunction ceases to have effect. If there are any particular passages which any of the parties consider ought to remain private, I will be prepared to consider submissions to that effect.