The competing arguments on appeal
82 The Government asserts that there are seven discreet errors involved in the primary judge's conclusion that the making of the 2018 ASA denied the respondent procedural fairness and was legally unreasonable. If the reference to discreet errors was intended to suggest that any one such error being established was sufficient to make good this ground of appeal, that may be doubted, not least because the relevant successfully appellable error must be more than something erroneous or infelicitous being detected even in some important aspect of her Honour's reasons. An error established in the reasons must be such as to vitiate the order that has been made. However, the better course in any event is to determine whether each such asserted error in the reasoning or conclusions has been established, and if so, then to determine whether the cumulative effect of any such errors is such as to produce vitiating error in the order setting aside the 2018 ASA.
83 In the greater part, the respondent does not directly respond to the Government's submissions, preferring instead to address the issues raised at a relatively high level of abstraction, largely divorced from the primary judge's reasons. In summary, the respondent seeks to apply the reasoning in A v Home Secretary in relation to the prohibition on the use in court proceedings of evidence obtained by torture to executive decision-making, without sufficiently engaging with the express distinction drawn by the House of Lords between the two.
84 The respondent submits that the Government is wrong to suggest that the common law condemnation of torture results only in a refusal to admit its fruits as evidence in a court of law, which itself overstates the Government's argument given the self-imposed restrictions in the Torture Policy and the Torture Procedure. The argument the respondent advances for blurring the distinction clearly and expressly drawn in A v Home Secretary, a decision that was endorsed in Shagang Shipping in 2020, is that the statutory context for the making of security assessments under s 37 of the ASIO Act is such that the use in any way of evidence obtained by torture is absolutely proscribed, relying upon the effect of a contrary conclusion to merits review by the Administrative Appeals Tribunal, which is itself an executive, not judicial, body.
85 No compelling argument is advanced by the respondent as to why this Court should depart from the approach taken in A v Home Secretary of finding that the common law allows the executive to make use of information obtained by a third party using torture, noting that ASIO requires such use to be cautious and to go to the question of weight. The respondent asserts that the making of a security assessment is ordinarily a more deliberative process than responding to a bomb threat, overlooking the wider ambit of executive decision-making also considered in A v Home Secretary. The respondent's argument also overlooks the fact that a security assessment may be just as urgent as any other executive decision, and an arrest decision or an investigative decision may be taken over a considerable period of time.
86 The period of time over which an executive decision may be able to be made is not a sound basis for deciding what may or may not be taken into account in making such a decision. Further, as was pointed out by Lord Rodger in A v Home Secretary at [135], an executive decision made using information obtained by a third party using torture will not necessarily be short-term, because the certificate of the Secretary of State was not required to be reviewed for six months, which was not short-lived or preliminary.
87 Turning to the detail of the Government's arguments, they first contend that the primary judge erred in finding that the reliance on evidence from the Returnees from Albania trial was "material and significant" (J [8(a)]), and was used and relied upon in a "primary and material way" adversely to the respondent (J [337]). The basis for that contention is that, when regard is had to the express acknowledgements in the 2018 TSOG of the cancellation of the Red Notice at [20] and of the criticisms of the Returnees from Albania trial at [29], some weight was placed on that information, noting that it is stated that the information sometimes accorded with other available information, and of the express statement at [48] that the information derived from that trial was being treated with caution. However, it is submitted, this did not meet the description given to it by her Honour of being material and significant, let alone being used in a primary and material way.
88 The respondent submits that this Court should not countenance a distinction between reliance that is material and significant and reliance that falls short of that, because the common law's condemnation of torture does not turn upon such fine gradations. This argument overlooks the fact that the primary judge relied upon that very characterisation as part of the reason for overturning the 2018 ASA. The respondent further submits that in any event the extent of the references to Returnees from Albania trial material in the 2018 TSOG meant that the primary judge was correct to characterise the use as material and significant.
89 An evaluative assessment of this kind by a primary judge is not always easy to counter, let alone characterise as erroneous. However, in this case it is difficult to see how the Director-General's reliance on the Returnees from Albania trial material fairly meets the description of being both material and significant, let alone that it was used in a primary and material way, given how it was said in the 2018 TSOG to be, in effect, secondary by giving it some weight, but merely contributing to the broader intelligence case underlying the security assessment. That said, a possible degree of overstatement alone would not be enough to make good this ground of appeal. However, it does feed into the Government's second asserted error.
90 Secondly, the Government contends that the primary judge erroneously, at J [345]-[354], concluded that any reliance on evidence from the Returnees from Albania trial was contrary to ASIO's policies. That does seem to be what her Honour found, albeit in support of the conclusion already reached. The Government relies upon the terms of [7.2.2] and [7.2.4] of SAD No. 2 reproduced above at [27], being part of a legislatively authorised determination for the purposes of the ASA aspect of the ASIO Act, and upon the terms of the Policy and of the Procedure, submitting, in substance that the SAD No. 2 and the related Policy and Procedure required caution, but not prohibition, in the use that was put to information that may have been derived from torture or CIDTP by others. The Government relies upon the following features of SAD No. 2 and related submissions:
(a) Paragraph 7.2.2 of SAD No. 2 provided that a security assessor should consider (a) the "credibility, nature and authenticity of the relevant facts, information and sources"; and (b) the weight to be accorded to the available information, "including whether the information can be corroborated";
(b) Paragraph 7.2.4 of SAD No. 2 provided that the weight to be given to information that "may be affected by the risk that it has been obtained by means which may amount to duress, torture or other cruel, inhuman or degrading treatment or punishment" and that in deciding the weight to be given to such information, should apply the Torture Policy;
(c) Contrary to the finding of the primary judge, no policy required ASIO to refuse to take into account evidence from the Returnees from Albania trial, but rather that the relevant parts of the Torture Policy are:
(i) from the Principles:
ASIO does not act in a way that sanctions, acquiesces to, or encourages torture or other cruel, inhuman or degrading treatment or punishment by others.
(ii) from the intent of the Policy:
This policy and the related Intelligence Procedure must be considered and, where appropriate, implemented in any of ASIO's activities, including (but not limited to) interviews, human source operations, special intelligence operations and collaboration activities.
The related Intelligence Procedure provides guidance on steps to be taken when an ASIO staff member becomes aware that information ASIO has received has been, or is assessed as likely to have been, derived from the use of torture or other CIDTP.
(d) The Torture Procedure requires:
(i) a staff member to:
Take into account in any assessment of the reliability and credibility of the information that the information has been, or is likely to have been, obtained by the use of torture or other CIDTP, and take suitable caution in the further use or dissemination of the information.
and
(ii) ASIO, when it becomes aware that information or intelligence it has received has been, or is assessed as being likely to have been obtained in that way, "assessments made in respect of the reliability and credibility of the information must be informed by that fact".
91 The substance of the Government's submission is that the 2018 TSOG reflects compliance with, rather than departure from, the above. Implicit in this submission is that the primary judge's findings that amount to attributing a difference in approach to that stated in those reasons is a form of merits review.
92 The Government's argument in relation to the primary judge having erred in finding a breach of policy by any reliance on evidence obtained from the Returnees from Albania trial must be accepted. An important but fine distinction is to be drawn between:
(a) ASIO itself engaging in torture or CIDTP, which is proscribed in Australia by Div 274 of the Criminal Code (Cth);
(b) ASIO doing anything to encourage such conduct by others, which is proscribed by ASIO's policy as detailed above; and
(c) restraint and caution by ASIO in using information that may (but does not have to be shown was) obtained by such conduct by others, which is otherwise not prohibited, relevantly for the purpose of conducting a security assessment and especially for making an ASA, with the Torture Policy in substance treating the risk of torture in this limited circumstance as going to reliability and thereby weight, rather than being a wholly impermissible source of information.
93 With the greatest of respect to the primary judge, her Honour appears to have conflated the criminal prohibition on ASIO itself engaging in conduct of that nature, or on ASIO in some way endorsing such conduct by others which is proscribed at least as a matter of policy, and the use of information obtained by others engaging in such conduct, and it then coming into the hands of ASIO, falling short of any such complicity, which is not prohibited by law or policy, but is required to be treated with restraint and caution. In doing so, her Honour erred. Doubtless even the deployment of information that may have been obtained in that way is troubling, and best avoided as Mr Cornall recommended should happen for any future security assessment of the respondent, and as was accepted for the process that gave rise to the 2020 ASA, but it is not absolutely forbidden when a question of national security is at stake and possible serious risks to this country and its people need to be assessed.
94 Thirdly, the Government contends that the primary judge at J [323] and [326] erroneously proceeded upon a misapprehension that a submission made at trial was that evidence potentially obtained by others by torture could only be used when it was corroborated, and that no such submission was made. Rather, consistently with the case outlined as addressed above in relation to the second asserted error, the Government's case advanced before her Honour was in written closing submissions at [85], with emphasis added in the Government's written submissions to this Court:
To the extent the validity of the 2018 ASA arises, the 2018 TSOG made clear at [20] (ECB 2450 (Vol B pdf p 1839)) that while some weight had been given to the evidence from the Returnees from Albania trial, in some cases that evidence corresponded with other evidence available to ASIO and in other cases to information provided by the applicant himself, and that it "merely contributes to the broader intelligence case underlying this security assessment".
95 Paragraph 20 of the 2018 TSOG states (omitting footnotes):
In February 2018, ASIO received information from the AFP that on 9 February 2018, the Commission for the Control of Interpol Files (CCF) had cancelled the IRN relating to [the respondent]. Information provided by the AFP indicates the review of the IRN relating to [the respondent] was undertaken after submissions were made to the CCF by [the respondent] regarding the purpose for which the IRN was being used, as well as the veracity of the charges that underpinned the IRN. Correspondence from the CCF, provided to ASIO by Home Affairs, indicated that after an examination of the relevant data the CCF found there were concerns relating to the data's compliance with Interpol's applicable rules and as a result, the information related to [the respondent] had been deleted from Interpol systems.
96 Two points may be made about this argument. First, it is by no means clear that the primary judge did proceed upon any misapprehension that the Government was submitting that evidence potentially obtained by torture could only be used when it was corroborated. Neither J [323] nor J [326], either individually, or in combination, say that. Secondly, even if her Honour had reached such a conclusion, it is far from clear that the substance of what was submitted by the Government could not fairly be read in that way. The totality of the Government's written submission at [85] reproduced above is open to being read in that way by referring to evidence that either corresponded with other evidence, or with information provided by the respondent, which amounts to saying that there was some form of corroboration. It follows that this third asserted error is not made out.
97 Fourth, the Government submits that the primary judge erred in her Honour's treatment of the CCF Report by treating aspects of the substantial extracts reproduced at J [326] as findings, when significant portions were in fact submissions by the respondent. It is submitted that this then formed the basis for the conclusion that it was "absolutely clear" that ASIO should not rely on evidence obtained from the Returnees from Albania trial, and that the 2018 TSOG represented a "sanitised summary" (J [330]), glossing over and omitting matters (J [331]), and a "fundamentally inaccurate" downplaying and diminishing of the findings from that report (J [338]), referring to J [328]-[331] and [337]-[338]).
98 Some aspects of this argument are better made than others. As to the treatment of the quoted passages from the CCF Report reproduced at J [326] as findings, while it is true that parts of those passages were indeed submissions made by the respondent as is clear from the text that was quoted, each such submission was accompanied by references to material apparently relied upon by him in making those submissions. Further, parts of what was reproduced were the views of international human rights organisations. Moreover, the Government's argument that the actual findings by the CCF in its report were much more limited fail to acknowledge that her Honour reproduced those parts as well.
99 The Government's criticism of the primary judge's conclusions drawn from the CCF Report, and of its treatment in the 2018 TSOG are better made. First, the CCF Report reflects a cautious approach to the use of information which may be too unreliable by reason of torture for something as serious as a Red Notice, which could result in someone being arrested and detained by a police force and then subject to extradition processes and court proceedings. The report was reflective of standards of data integrity and reliability that the Interpol Constitution requires to be met before the potential use of law enforcement powers is triggered, which in turn is likely to result in curial processes. That is quite a different framework to intelligence and risk assessment, in particular directed to questions of, and concern with, national security. It is another way of looking at the distinction between use of evidence in court, and use of material for executive decision-making.
100 Thus, when the primary judge at J [327] suggested that the reasons for the CCF deciding to cancel the Red Notice, which is not attributed to caution given the nature of the law enforcement powers that it could trigger, but rather to more generalised criticism of the trial process, meant that "no officer of any Australian government agency acting reasonably could have failed to understand that evidence from these sources should not be used", her Honour conflated two quite different uses, objectives and purposes. The views that her Honour expressed may well be apposite for an Australian government agency charged with law enforcement responsibilities ordinarily directed to potentially initiating curial processes, such as the Australian Federal Police, which is not what this case concerns. But that does not readily equate to the protective intelligence activities and responsibilities of ASIO and like agencies in the intelligence community. There is a material difference between using such information in aid of law enforcement potentially ending in a curial process being brought against a person by the State, and it being used, albeit cautiously, as part of intelligence and thereby risk assessment processes, including culminating in an adverse security assessment. As the discussion about this topic below makes clear, this distinction is well-recognised.
101 Secondly, the primary judge's criticisms of the 2018 TSOG's use of the Returnees from Albania trial information are similarly misplaced. Once the misplaced premise that there was any complete policy prohibition on the executive using information that bears the risk of having been obtained by torture or CIDTP is put to one side, and it is appreciated that the requirement to treat such material with caution was in fact observed, then the use to which information from the Returnees from Albania trial was put can be better and more appropriately understood. The 2018 TSOG is a redacted version of the reasons for the making of the 2018 ASA, and in places quite heavily redacted both in the text and in the footnotes. Read fairly, the Returnees from Albania trial information forms a relatively small part of what was taken into account, and was far from dominant, determinative or even especially significant, at least in the unredacted parts able to be seen by the primary judge and by this Court. The description in the 2018 TSOG of it being used in a supportive way is accurate. To describe it as material and significant entails a measure of overstatement, but to describe it as both primary and material is unsustainable.
102 The primary judge gave little overt consideration to the bulk of the 2018 TSOG, placing most of the determinative weight on its treatment of the CCF Report. In particular, the primary judge at J [330]-[331] said that the last sentence in [20] of the 2018 TSOG (reproduced above at [95]) - "Correspondence from the CCF, provided to ASIO by Home Affairs, indicated that after an examination of the relevant data the CCF found there were concerns relating to the data's compliance with Interpol's applicable rules and as a result, the information related to [the respondent] had been deleted from Interpol systems" - was "somewhat sanitised", and further characterised it as glossing over and omitting the parts that her Honour had quoted at J [326]. Yet that impugned sentence is an entirely accurate summary of the ultimate conclusion that the CCF Report reached. The CCF reproduced and acknowledged the criticisms that had been made of the Returnees from Albania trial, as quoted in part by her Honour at J [326], and found that the Egyptian branch of Interpol, who had caused the Red Notice to issue, had not satisfactorily answered those concerns sufficiently for the CCF to be satisfied that the data giving rise to the issuing of the Red Notice met the necessary standard of reliability.
103 The CCF did not find, one way or the other, that the criticisms of the Returnees from Albania trial, including as to obtaining evidence by torture, had in fact been established. The apparently simple reason for that which is evident from a careful consideration of the entirety of the CCF Report is that, in effect, the requirement placed on the Egyptian branch of Interpol was to show that these allegations were not true so that the information upon which the Red Notice was based met law enforcement reliability standards, not for the respondent to show that they were true. Such a stance is inherently appropriate for that purpose, placing a precautionary barrier in the way of relying upon material for which there is a real risk that it has been obtained by the use of torture, resulting in the issue of a Red Notice. The 2018 TSOG was recording the conclusion ultimately reached by the CCF accurately, and did not repeat the allegations leading to that conclusion which were not in fact found to be true, because they did not need to be.
104 It follows that the Government has made good the assertion of error on the part of the primary judge in overstating the effect of the CCF Report, and then using that characterisation to impugn any use of material from the Returnees from Albania trial and thereby conclude that this information could not be used at all. There was no such error, let alone vitiating error, in having regard to the Returnees from Albania trial material in the limited way it was deployed.
105 The Government's fifth argument flows from the fourth, infecting the conclusions reached by the primary judge at J [337] and J [341]-[342] about use of the material obtained as part of the Returnees from Albania trial. Her Honour said:
[337] It suffices at this point to make clear that, in my opinion, the applicant has proven a denial of procedural fairness by ASIO in the making of the 2018 ASA, because of the way that ASIO used and relied upon, in a primary and material way and adversely to the applicant, evidence obtained as part of the Returnees from Albania trial that had been thoroughly and completely discredited by a range of international organisations and foreign courts, most importantly Interpol's CCF, being one of the very international organisations whose IRN both ASIO and the AFP (and other agencies such as Home Affairs) had acted upon in their former treatment of, and decision-making about, the applicant.
…
[341] Unlike Jaffarie and SCDV, this is not a denial of procedural fairness by way of a failure to disclose sufficient information that could and ought to have been disclosed. This is a denial of procedural fairness by reliance on material that could not be treated as credible, reliable and probative. It was a process that was unfair to the applicant: see Minister for Immigration and Border Protection v WZARH [2015] HCA 40; 256 CLR 326 at [38], Kiefel, Bell and Keane JJ. It was not "a procedure which conforms to the procedure which a reasonable and fair repository of the power would adopt in the circumstances": WZARH at [56], Gageler and Gordon JJ, quoting Kioa v West [1985] HCA 81; 159 CLR 550 at 627, Brennan J.
[342] There is no difficulty in describing what happened as a practical injustice to the applicant. The evidence relied on in the 2018 ASA from the Returnees from Albania trial had been so thoroughly discredited, and known to be so, that it should never have been relied on by ASIO. It should never have been sought to be justified to Mr Lewis who, given his position, could not have been expected to go and check what he was being told in the briefing note. The subsequent view taken by Mr Cornall confirms this, as does ASIO's own change of position in 2020. Because of this reliance, the applicant was denied a fair consideration of his own evidence and answers, especially about the matters touched on at [103]-[106] of the 2018 TSOG, which were central to ASIO's concerns and remained so throughout the three assessments. Instead, his evidence and answers were measured against evidence which had been found to be tainted by the likelihood of its extraction by torture, and/or because it was "pre-prepared" by Egyptian authorities.
106 The Government submits that in addition to the factual basis for those conclusions being absent (being the fourth argument considered above) there is no authority cited by either the primary judge or the respondent for the proposition that relying upon material that is not credible or reliable is procedurally unfair, with WZARH not being such an authority because that decision at [38] refers to denial of an opportunity to be heard. The Government submits that, given the cautious approach taken with the evidence from the Returnees from Albania trial, there was no procedural unfairness in the process adopted. ASIO had weighed that evidence against the respondent's evidence, and the use of this evidence did not deny him a meaningful opportunity to participate in his security assessment, especially given that he had been interviewed and invited to comment on four occasions. Moreover, it is submitted, her Honour had earlier accepted at J [213] that the respondent "well understands ASIO's reasoning at least in relation to its findings about his past activities and associations, even if he does not agree with them, and even if he does not know all the factual sources from which that reasoning is said to derive".
107 The Government's argument on the finding of a denial of procedural fairness must be accepted. There was no such denial.
108 The Government's sixth argument, again flowing from the preceding arguments, and in particular the fourth argument, is that the primary judge erred at J [327] and J [340] where her Honour said:
[327] Just as, quite properly, Australian authorities including the AFP and ASIO took the existence of an IRN as a serious matter against the applicant's interests, it would have been irrational, and legally unreasonable, for the same Australian authorities not to give careful and serious consideration and respect to the decision of the same organisation to cancel the IRN, including careful and serious consideration of why it had decided to do so. The passages extracted above record the widespread condemnation of the Returnees from Albania trial, and the methods used to extract evidence. After Interpol's CCF report at the very least, no officer of an Australian government agency acting reasonably could have failed to understand that evidence from these sources should not be used.
…
[340] In taking this course in the 2018 assessment, ASIO officers consciously and actively relied on information from sources any officers acting rationally and reasonably would have discarded as not only wholly unreliable, but as against public policy for an Australian government agency such as ASIO to use. That such policy existed is evident in cl 7.2.4 of SAD2, as extracted above. That position is consistent with Commonwealth legislation. Section 27D of the Foreign Evidence Act 1994 (Cth) prohibits the admission of "foreign material" or "foreign government material" in a "terrorism-related proceeding" if the court is satisfied that the material, or information contained in the material, "was obtained directly as a result of torture or duress" (as defined in that provision). It is also consistent with judicial approaches to this kind of evidence or material. In A v Secretary of State for the Home Department (No 2) [2005] UKHL 71; [2006] 2 AC 221 at [52], Lord Bingham of Cornhill said:
The principles of the common law, standing alone, in my opinion compel the exclusion of third party torture evidence as unreliable, unfair, offensive to ordinary standards of humanity and decency and incompatible with the principles which should animate a tribunal seeking to administer justice.
109 The Government submits that it was not open for her Honour to conclude that once it was sufficiently clear that evidence was tainted by torture, it was legally unreasonable for ASIO to use that information in any way, and again that this was contrary to public policy as reflected in SAD No. 2. This argument must also be accepted in light of the conclusions already reached to the effect that there was no legal reason, nor public policy imperative, nor ASIO policy dictating that regard could not be had to the material from the Returnees from Albania trial, especially in the limited and cautious way in which it was deployed. The error was compounded by reliance upon prohibitions that only apply to adducing evidence in court proceedings that has been shown to be obtained by the use of torture, which is the limitation that both s 27D of the Foreign Evidence Act, and Lord Bingham in A v Home Secretary at [52] refer to and are confined to.
110 The seventh and final argument advanced by the Government is that the primary judge erred at J [342] in concluding that the Cornall Report confirms that the evidence from the Returnees from Albania trial should never have been relied upon. Strictly speaking, this point goes nowhere. However, as the Government correctly points out, Mr Cornall in fact found, as one of three reservations about the 2018 ASA, that the Returnees from Albania trial was a travesty of justice and that any reliance upon the so-called evidence and judgment in that case, and the respondent's conviction in absentia, detracted from the quality of the security assessment, further noting that there was in any event very limited information in the Egyptian judgment about the respondent, or in the statements against him said to have been obtained by torture. As the Government correctly points out, Mr Cornall did not find that the reliance upon that material met the description of being irrational, unreasonable, unfair, or contrary to any ASIO policy or procedure. The substance of the criticism made by Mr Cornall is that the evidence from that source was not really worth using.