Dr Sam Raphael v Information Commissioner & Anor

The request relates to actions taken pursuant to MOD Policy on the Passing or Receipt of Intelligence Relating to Detained or Captured Persons, the latest version of which was released to me on 13 March 2019 (F012019/01980). It also relates to all earlier versions of this policy document (e. g. the May 2013 version, released to me on 8 December 2014 (F012014/05808)).

Please provide me with overall figures, broken down by year, for the following:

Number of times that Ministers have been consulted in cases where MOD officials considered there to be a serious risk of torture and/or CIDT which cannot be mitigated (para 15 (viii) and (ix));

Number of times where Ministers have been consulted as above, and have subsequently approved intelligence sharing;

Number of times that prior approval has been sought from Ministers (para 22)

Number of times where Ministers have provided prior approval as above.

Similar figures were released to me on 8 December 2014, for the two years 2013 and 2014 (F012014/05808). The information I am seeking here is identical in nature to this. '

In July 2010, the Cabinet Office published Consolidated Guidance to Intelligence Officers and Service Personnel on the Detention and Interviewing of Detainees Overseas, and on the Passing and Receipt of Intelligence Relating to Detainees (“the Consolidated Guidance”) The Consolidated Guidance provided that before interviewing or seeking intelligence from detainees in the custody of an overseas security or intelligence service, or seeking an individual’s detention by such a service, UK personnel were required to consider whether that person “may have been or may be subjected to unacceptable standards of detention or treatment” (§9). The Guidance set out different actions to be followed depending on the degree of identified risk of such detention or treatment (in the “Consolidated Guidance Table”, §11. The Consolidated Guidance Table materially provided:-

THE LAW

23.— Information supplied by, or relating to, bodies dealing with security matters.

Information held by a public authority is exempt information if it was directly or indirectly supplied to the public authority by, or relates to, any of the bodies specified in subsection (3).

Section 23 affords the “widest protection” of any of the exemptions: Cobain at [19(b)] and [29].

The purpose of section 23 is to preserve the operational secrecy necessary for section 23(3) bodies to function: Lownie at [50].

It is “Parliament’s clear intention that, because of what they do, there should be no question of using FOIA to obtain information from or about the activities of section 23 bodies at all”. The exclusion of the section 23(3) bodies from the scope of FOIA was shutting the front door, and section 23 was “a means of shutting the back door to ensure that this exclusion was not circumvented”: APPGER at [16].

The legislative choice of Parliament was that “the exclusionary principle was so fundamental when considering information touching the specified bodies, that even perfectly harmless disclosure would only be made on the initiative or with the consent of the body concerned”: Cobain at [28]; Lownie at [53].

Asking whether the information requested is anodyne or revelatory fails to respect the difficulty of identifying what the revelatory nature of the information might be without a detailed understanding of the security context: Lownie at [42]; Corderoy at [59].

When applying the ‘relates to’ limb of sections 23(1) and (5), that language is used in “a wide sense”: APPGER at [25]; Corderoy at [59]; Savic at [40].

The first port of call should always be the statutory language without any judicial gloss: APPGER at [23]; Corderoy at [51]; Savic at [40].

With that warning in mind, in the context of ‘relates to’ in section 23, it may sometimes be helpful to consider the synonyms of “some connection”, or “that it touches or stands in some relation to” (APPGER at [13], [25]) or to consider whether the request is for “information, in a record supplied to one or more of the section 23 bodies, which was for the purpose of the discharge of their statutory functions” (APPGER at [21], [26]; Lownie at [57]). But the ‘relates to’ limb must not be read as subject to a test of focus (APPGER at [14) or directness (Lownie at [59]- [60]).

The scope of the ‘relates to’ limb is not unlimited and there will come a point when any connection between the information and the section 23(3) body is too remote. Assessing this is a question of judgment on the evidence: Lownie at [62].

The assessment of the degree of relationship may be informed by the context of the information: Lownie at [4] and [67].

The scope of the section 23 exemption is not to be construed or applied by reference to other exemptions, including section 24: APPGER at [17]; Lownie at [45] and [52].

In a section 23(1) case, regard should be had as to whether or not information can be disaggregated from the exempt information so as to render it non-exempt and still be provided in an intelligible form: Corderoy at [43].

24.— National security.

Information which does not fall within section 23(1) is exempt information if exemption from section 1(1)(b) is required for the purpose of safeguarding national security.”

The term national security has been interpreted broadly and encompasses the security of the United Kingdom and its people, the protection of democracy and the legal and constitutional systems of the state: Secretary of State for the Home Department v Rehman [2001] UKHL 47, [2003] 1 AC 153, paras 15-16 per Lord Steyn, para 50 per Lord Hoffmann and para 64 per Lord Hutton.

A threat to national security may be direct (the threat of action against the United Kingdom) or indirect (arising from the threat of action directed against other states): Rehman, paras 16 and 64.

Section 24 is not engaged, unlike the majority of the qualified exemptions, by a consideration of prejudice. Its engagement is deliberately differently worded.

The term “required” means “reasonably necessary”: Kalman v Information Commissioner & Department for Transport [2011] 1 Info LR 664, para 33.

National security is a matter of vital national importance in which the Tribunal should pause and reflect very carefully before overriding the sincerely held views of relevant public authorities: APPGER v Information Commissioner & Ministry of Defence [2011] UKUT 153 (AAC), [2011] 2 Info LR 75, para 56 (citing Rehman).

Even where the chance of a particular harm occurring is relatively low, the seriousness of the consequences (the nature of the risk) can nonetheless mean that the public interest in avoiding that risk is very strong: Kalman, para 47. As the Upper Tribunal put it: “the reality is that the public interest in maintaining the qualified national security exemption in section 24(1) is likely to be substantial and to require a compelling competing public interest to equal or outweigh it”: Keane v Information Commissioner, Home Office and Metropolitan Police Service [2016] UKUT 461 (AAC), para 58 (approving Kalman). That does not mean that the section 24 exemption carries “inherent weight”, but is rather a reflection of what is likely to be a fair recognition of the public interests involved in the particular circumstances of a case in which section 24 is properly engaged.”

This broad approach by reference to identified bodies [in section 23] is not narrowed by the qualified exemption in section 24(1), namely that information which does not fall within section 23(1) is exempt information if exemption from section 1(1)(b) is required for the purpose of safeguarding national security. This is a safety net provision which recognises that national security issues may arise in respect of information that is not within the absolute section 23 exemption. Rather this safety net provision reinforces the view that Parliament's intention was to put section 23 bodies outside the ambit of the right to information conferred by FOIA and a narrow approach to an absolute exemption would not promote that purpose.

27 (1) Information is exempt information if its disclosure under this Act would, or would be likely to, prejudice—

….

…

The duty to confirm or deny does not arise if, or to the extent that, compliance with section 1(1)(a)—

There are essentially two issues:

would disclosure of the information be likely to prejudice international relations;

if so, does the public interest in maintaining the exemption outweigh the public interest in disclosing it.

Both matters are for the Tribunal to determine for itself in the light of the evidence. Appropriate weight needs to be attached to evidence from the executive branch of government about the prejudice likely to be caused to particular relations by disclosure of particular information: see Secretary of State for the Home Department v Rehman [2001] UKHL 47; [2003] 1 AC 153, [50]-[53] and see also R (on the application of Mohamed) v Secretary of State for Foreign and Commonwealth Affairs [2010] EWCA Civ 65 at [131] per Master of the Rolls:

“In practical terms, the Foreign Secretary has unrestricted access to full and open advice from his experienced advisers, both in the Foreign Office and the intelligence services. He is accordingly far better informed, as well as having far more relevant experience, than any judge, for the purpose of assessing the likely attitude and actions of foreign intelligence services as a result of the publication of the redacted paragraphs, and the consequences of any such actions so far as the prevention of terrorism in this country is concerned.”

…the MOD explained to the Commissioner that intelligence sharing agreements, and information relating to any exchanges made under that theme, are based on mutual trust and considered confidential between the relevant parties. The MOD argued that whilst it is public knowledge that there is a UK and MOD policy which outlines the guidelines that have to be followed when passing or receiving intelligence relating to detained or captured persons, the expectation is that information relating to any exchanges considered under that policy is not revealed publicly by the parties involved. Therefore, the release of such information would be considered a breach of trust that would likely to prejudice the UK’s relationships with the foreign authorities that shared the intelligence for which Ministerial advice was sought. However, the MOD noted that any perceived UK breach of trust relating to the sharing of intelligence could have wider implications for the UK’s relations with other states which involve the sharing of sensitive or classified material.

As explained below, the complainant questioned why the MOD had sought to withhold this information when it had disclosed substantively the same information in response to a previous request five years earlier. The complainant argued that the MOD had failed to explain what circumstances had changed over the period.

In response to this point the MOD explained to the Commissioner that during 2013 and 2014 (the period covered by the complainant’s previous request) the UK Armed Forces were actively engaged in ground operations in Afghanistan under Operation HERRICK. The MOD explained that during this operation, UK forces had been conducting detention operations that involved the transfer of captured persons to Afghan authorities, including Afghan police forces where criminal acts falling under their jurisdiction were believed to have been committed. The MOD explained that the NATO International Security Assistance Force (ISAF) set a deadline of December 2014 to end combat operations in Afghanistan, and in line with this the UK’s role gradually shifted from one of combat to training and assistance over that period, with the formal withdrawal of UK combat forces in late 2014. During this transition, responsibility for security passed from ISAF to Afghan National Security Forces (ANSF). The MOD explained that this is a material change to the UK Armed Forces operating environment between 2013-2014 and the present time.

However, the MOD explained that it was now of the view that the information released in response to the complainant’s previous request should not have been disclosed. It explained that the publication of information that provides the public and adversaries with an insight into the MOD’s intelligence sharing with partner forces and highly sensitive UK operational data was an error, and one that should not be repeated.

The MOD explained that the approach taken in this request was supported the position outlined in the written statement by the Minister for the Armed Forces given on 11 June 2019:

14….he had requested substantively the same information as released in a previous request (FOI2014/05808), but the MOD had not explained what had changed in the intervening five years which meant the such information was now considered to be exempt from disclosure. The complainant suggested that it seemed possible that the only change is the greater public scrutiny on intelligence sharing following the release of the MOD's policy issue on this as a request of a separate request he had made for it.

The complainant noted that the MOD had argued that the information is ‘operationally sensitive’ and would ‘prejudice the capability or effectiveness of our armed forces’. He also noted that the MOD had argued that it would also ‘be likely to adversely affect relations with our allies if revealed.’ However, the complainant argued that it was very unclear how the information requested would produce this harm. He emphasised that he had requested aggregate figures, which the MOD had released to him before; he also emphasised that he did not ask for details of the nature of intelligence to be shared, nor the identity of the ally. Furthermore, he noted that the refusal notice argued that prejudice would be likely to occur in combination with ‘other information that could be revealed under the FOIA’, but it is unclear what that information would be and how it would combine to generate harm. He suggested for this argument to hold the likelihood of that harm occurring would need to be realistic not purely hypothetical and the harm genuine.

18.….the Commissioner accepts that the potential prejudice described by the MOD clearly relates to the interests which the exemption contained at section 27(1)(a) is designed to protect.

….the Commissioner acknowledges that there is an expectation that information shared between states on the basis of intelligence sharing agreements will be treated confidentially. In light of this she accepts that it is plausible to argue that disclosure of the withheld information which relates directly to information provided to the UK under such an agreement would be against the expectations of the states that provided it. In turn, she accepts that disclosure of the information would therefore have a negative impact on the UK’s relations with those states, or as described by the Tribunal above, would require a damage limitation response that would otherwise have not been necessary.

….the Commissioner is persuaded that the risk of prejudice occurring is one that is more than hypothetical. In reaching this conclusion she appreciates that the information requested is simply aggregate data and does not identify which states provided the data. However, given the clear expectation of the states in terms of how information shared with the UK would be treated, and the inherent sensitivity of such information, the Commissioner is satisfied that even disclosure of the such aggregate data still poses a real and significant risk to the UK’s relations with the states who provided it with the intelligence.

The Commissioner accepts that there is a significant public interest in the disclosure of the information sought by the complainant. As he noted, following the MOD’s disclosure of its policy regarding such intelligence there were Parliamentary debates and media articles about this matter, including questions being raised as to whether the policy was compliant with the Cabinet Office’s consolidated guidance on torture. Disclosure of the withheld information would provide the public with a clear insight into the number of occasions Ministers were consulted and prior approval had been sought for a five year period. The Commissioner accepts that disclosure of the information could further inform ongoing debate on the operation of the policy in question.

However, the Commissioner considers there to be a very significant public interest in protecting the UK’s relationships with other states, not least to ensure that intelligence sharing arrangements are not compromised. In attributing weight to the public interest in favour of maintaining the exemption the Commissioner accepts that the disclosure risks not only directly harming the UK’s relations with the states that provided the intelligence in question but also risks undermining the UK’s intelligence sharing relations with other states. In the Commissioner’s view this adds further weight to the public interest in maintaining the exemption.

The question now arising for determination by the Tribunal is therefore whether the MOD is entitled to withhold information relating to the number of times Ministers are consulted on, and approve, information sharing in circumstances in which there is a serious and unmitigable risk that such action will lead to detainees being tortured and/or subjected to CIDT, on the basis of one or more of: sections 23(1), in the alternative section 24(1); section 26(1)(b); and/or section 27(1)(a).

Having corrected its earlier erroneous reliance on section 23(5) and section 24(1)FOIA (as to which see above), the MOD relies on section 23(1), which is an absolute exemption. Even if the MOD were unable to establish prejudice (which, including as set out in this statement, it can), the information must nevertheless be withheld.

As to section 24(1), withholding the requested information is necessary for the purpose of safeguarding national security. While it is public knowledge MOD has a policy to share and receive intelligence with foreign partners, disclosure of the requested information could have harmful impact on co-operation with other states, which could then impact on the security of the UK and its people. Release could undermine any ongoing or future operations and it is possible that the accumulation of information could enable interested parties to build a picture of UK intelligence sharing and how regularly it does, or does not occur, which is extremely sensitive to national security. Again, this information, when combined with other information in the public domain, such as regions of operations and known international partners, would create a mosaic of understanding and so risk national security.

Disclosing the requested information would undermine the UK’s position and compromise its ability to work with other states on matters of intelligence sharing relating to detainees or captured persons. As identified above, the exchange of intelligence between states is one of the most sensitive elements of any relationship and it is dependent on mutual trust. Intelligence is shared on the basis that such activities will remain confidential and not be disclosed. As a long-standing matter of policy, HMG does not comment on intelligence matters for reasons including the preservation of intelligence sharing relationships, and the same considerations inform the MOD’s reliance on section 27(1)(a) FOIA in this appeal.

Disclosing the requested information would prejudice the capability, effectiveness or security of UK Armed Forces and any partner forces they are operating with. The disclosure of the requested information would provide adversaries with a clear indication of the tempo, regularity and scale of level of intelligence sharing that may, or may not, have occurred with foreign authorities. This would contribute to the cumulative mosaic understanding that adversaries can build of the way in which sensitive military operations are conducted by British Armed Forces or our allies worldwide. This could in turn impact the security and protection of UK and partner forces.

While it is recognised that the Information Requests ask for numerical data in respect of certain actions, providing such numbers, even if zero, when collated with other information in the public domain or through FOIA requests, could reveal how active we may be in certain parts of the world and risk force protection issues if disclosed. As a consequence, this may prejudice our relationships with allies and partner forces, by adversely impacting on their perception of and willingness to engage with UK Armed Forces. Were foreign forces to limit their engagement with UK Armed Forces, that could result in a wider impact on the UK Armed Forces’ sensitive operational intelligence sharing relationships, which would be detrimental to UK Armed Forces capability.

While the release of the information might be seen to demonstrate the MOD’s commitment to transparency and openness regarding the conduct of military operations, the overall public interest is best served in withholding the requested information, given that the disclosure would prejudice the capability and effectiveness of UK Armed Forces and partner forces.

…In this OPEN statement, I am unable to address in substance the risk or potential risk arising from the disclosure of the requested information when considered in conjunction with other information publicly available or available to adversaries.

Nevertheless, I note that the erroneous prior disclosure of numerical information in the 2014 Response places a particular complexion on mosaic risk in this appeal. In 2014, information was disclosed concerning details of intelligence sharing in 2013 and 2014. Were the requested information to be disclosed, information reflecting a total of seven years of intelligence sharing would be placed in the public domain. That information would be capable of consideration alongside other information and reports regarding the activities of the UK Government and its forces, as well as foreign forces and allies, so as to illuminate and render public aspects of properly confidential intelligence sharing arrangements. Taken together with the 2014 Response, disclosure of the requested information would place a sizable and thereby conspicuous body of information in the public domain concerning intelligence sharing. In my view, that would or would be likely to undermine the UK Government’s relations with other states, prejudice the capability, effectiveness or security of UK Armed Forces and/or relevant cooperating forces, and prejudice national security.

The totality of the precise information held by adversaries is not known to HMG, unsurprisingly. The suggestion …that mosaic risk for the purposes of this appeal must be evidenced “by reference to specific information in the public domain” places an impractical and unduly onerous burden on the MOD, in my view.

While it is public knowledge that there are MOD policies which outline the guidelines that have to be followed when passing or receiving intelligence relating to detained or captured persons, the expectation is that information relating to any exchanges considered under that policy is not to be revealed publicly by the parties involved. The release of such information would be considered a breach of trust that would likely to prejudice our relationships with the foreign authorities. It should also be noted that any perceived UK breach of trust relating to the sharing of intelligence could have wider implications for our relations with other states which involve the sharing of sensitive or classified material.

Accordingly, the public interest balance lies heavily in favour of withholding the requested information, disclosure of which would undermine the UK’s position and compromise its ability to work with other states on matters of intelligence sharing relating to detainees or captured persons.

Submissions

The public interest in knowing whether, and to what extent, Ministers authorise intelligence sharing in circumstances in which there is a clear risk of torture or CIDT to detainees.

The public interest in helping to understand whether the UK Government, its Ministers, and MOD personnel are complying with their obligations under international and domestic law.

The public interest in knowing whether the UK Government is complying with its repeatedly stated policy of refusing to participate in, solicit, encourage or condone the use of torture.

Mrs Jennifer Armstrong (“the witness”) adopted the CLOSED witness statement of Anthony McGee.

The witness addressed the engagement of s.23(1), alternatively 24(1) FOIA, on the facts.

The witness addressed the relevance of the fact that other information had previously been disclosed, including:

The witness addressed the likely reaction of international partners and adversaries if the disclosure of the requested information were to come to their attention.

The witness gave granular and specific examples of mosaic risk arising generally and as applicable to the specific content of the requested information.

The witness was asked each of the questions provided by the Appellant in advance of the CLOSED session. Of those:

The Respondents delivered short oral CLOSED closing submissions in light of the CLOSED evidence. Each of the Respondents submitted that s.23(1), alternatively s.24(1) FOIA were engaged. As applicable, the public interest balance fell in favour of withholding the requested information in respect of ss.24(1) and 27(1)(a) (both Respondents) and additionally s.26(1)(b) (MOD only). The Commissioner did not make submissions on s.26(1)(b). Such submissions as can be made in OPEN, including submissions as to the public interest balance, were reserved to the hearing on 14 June 2023.