Abdul Nacer Benbrika, an Algerian-Australian dual national, was found guilty of being a member of a terrorist organisation, of directing the activities of a terrorist organisation, and of possession of a thing connected with preparation for a terrorist act. He was sentenced to a 12 year non-parole term of imprisonment in 2009. His three-year continuing detention order (CDO) has expired. The Supreme Court of Victoria has today, on an application from the Commonwealth, imposed an extended supervision order. At the time of publishing, the Court has not published its written reasons. This note is not intended as an evaluation of the course of action taken by the Commonwealth; it is an assessment of Division 105A of the Criminal Code Act 1995 (Cth) as relevant to the different courses of action that were available to the Commonwealth. This note does not suggest the Commonwealth should have sought a more serious post-sentence order than it did.
"The Commonwealth has taken the strongest possible action available under law in accordance with the advice from all operational agencies involved in the matter, including AFP and Victoria Police” a spokesperson for the Attorney-General told The Guardian last week, announcing the Government would apply for an extended supervision order (ESO) instead of a second continuing detention order (CDO).
One may argue it is unexplained why the Government has suggested this case warrants its “strongest possible action” on its advice while downgrading its response in the absence of a legal necessity to do so. The Act enables the Commonwealth to apply for a CDO and instead be granted an ESO if the Court views the threshold for a CDO to not have been met. The implication from its statement is that the Government has formed a view that the subject ceases to qualify for a CDO, an implication confirmed by the Attorney-General in a separate statement released this afternoon that has today been affirmed by the Victorian Supreme Court.
The decision taken by the Attorney-General was presumably informed by legal advice. Equally, it appears that at least two “operational agencies” have advised that the subject poses reduced risk and therefore does not warrant a CDO. However, strictly “under law”, the "strongest possible action" would have been to apply for a CDO.
There is no statutory preclusion to applying for a second CDO (ss105A.7(6)). The sole entity that can determine what is "possible" "under law" in this case is a court. That the Commonwealth applied only for an ESO means it has not formally asked a court to determine the suitability of a second CDO. The Government has instead made its own assessment of the advice from "operational agencies" and other unidentified legal advice, which is both appropriate and authorised.
A court that receives an application for a CDO is not limited to imposing or rejecting the proposed CDO; it can instead impose an ESO (ss105A.6A(1)(b)). However, a court that receives an application for an ESO can only impose or reject an ESO (s105A.6A(2)). The latter applies to this case. The legal decision taken by the Government on its advice has in fact foreclosed “the strongest possible action available under law”, albeit this may be appropriate given the decision was taken “in accordance with the advice from all operational agencies involved in the matter”. Given the verbally relayed reasoning for the Victorian Supreme Court today imposing an ESO of one year, two years lesser than the three year ESO sought by the Attorney-General, there is no cause to expect that a decision to instead have applied for a CDO would have procured a different outcome. The Victorian Supreme Court decision shows the Commonwealth sought a greater post-sentence order than the Court viewed was justified.
There was no risk that a rejected CDO application would have resulted in the subject being released from custody without a post-sentence order, given the Court could have come to an alternative judgement of imposing an ESO.
There is no statutory obligation on the Attorney-General to ensure consultation with security agencies on an expiring CDO. The Attorney-General must only ensure "reasonable inquiries" are made when applying for one (ss105A.5(2A)(a)). Having received advice from "...all operational agencies involved in the matter, including AFP and Victoria Police" may or may not identify all agencies with which the Government consulted. Given the absence of a statutory obligation to consult on an expiring CDO, the statute offers no further indication on whether other departments or agencies, such as ASIO, were consulted.
That there is no precedent of the Commonwealth applying for a second CDO means an application for a second CDO is without common law. Given the position taken by the Government to not seek a CDO, it remains legally untested, especially given that the Court is not prevented from “having regard to any other matter the Court considers relevant” in determining whether to impose one (ss105A.6B(2)).
The Attorney-General or its legal representative must apply to a court for a review of a CDO each twelve months that a CDO is in force to sustain it (ss105A.10(1B)(b)). The Court can affirm a CDO if it is satisfied it could have imposed a CDO (ss105A.12(4)(a)) pursuant to the provisions for first imposing one (s105A.7).
That the Court did not revoke the CDO (ss105A.12(5AA)) twelve months ago means it was satisfied it could have imposed a CDO. This means the Court was satisfied "to a high degree of probability... that the offender poses an unacceptable risk of committing a serious Part 5.3 [terrorism] offence" and "that there is no less restrictive measure available under this Part that would be effective in preventing the unacceptable risk."
Whether CDOs are a proportionate mechanism respectful of human rights is a legitimate, albeit separate, question. Twelve months after the Government applied for a review which affirmed his CDO, the Government has decided against seeking another.
A consequential question is: what has changed?
Again, the confirmed implication is that the Government has formed a view that the subject ceases to qualify for a CDO, a view today affirmed by the Victorian Supreme Court.
The Sydney Morning Herald and The Age reports today that Justice Elizabeth Hollingworth said:
The evidence clearly establishes, and the Attorney-General quite clearly accepts, that Mr Benbrika has been making substantial progress towards de-radicalisation in recent years,
The risk of offending is now low enough that it can be managed by Mr Benbrika living in the community on a strict ESO.
Separately, the The Sydney Morning Herald and The Age yesterday published an allegation attributed to former Independent National Security Legislation Monitor Grant Donaldson SC:
Donaldson, who has previously called on the government to abolish continuing detention orders, said in a report last week that his stance was bolstered by a 2020 report by the Australian National University’s Dr Emily Corner, who found the tool used to keep Benbrika locked up could not accurately assess the risk.
The Department of Home Affairs commissioned the report, but it was never disclosed to Benbrika’s defence team.
The Act requires the Attorney-General to, during a review of a post-sentence order (s105A.12), "ensure that reasonable inquiries are made to ascertain any facts known to any Commonwealth law enforcement officer or intelligence or security officer that would reasonably be regarded as supporting a finding that the order should not be affirmed" (ss105A.12(5A)). The Attorney-General or its representative must "present to the Court... (a) a copy of any material in the possession of the AFP Minister or legal representative; and (b) a statement of any facts that the AFP Minister or legal representative is aware of; that would reasonably be regarded as supporting a finding that: (c) the order should not be affirmed, or should not be affirmed in the terms in which the order is made" (ss105A.12(6A)). The allegation attributed to Grant Donaldson SC gives rise to a question of whether the Commonwealth acted lawfully if the alleged Australian National University report commissioned by the Department of Home Affairs was not tendered in the two reviews held since the subject first had a CDO imposed.
19 December 2023
Popular encrypted email provider Proton released a “high-security program” intended for its users most at risk of targeted attacks. Proton claimed it “surpasses everything that has come before”. This assertion is not supported by high quality, longstanding evidence. Worse, Proton did not acknowledge some of its persistent security development inferiority that exposes its users to otherwise preventable attacks, the same attacks its new program claims to counter, but that it will not prevent.
“The Proton Sentinel program distinguishes itself from other enhanced protection programs by going beyond strengthening the default protections (Proton’s defaults are already very secure). Proton Sentinel surpasses everything that has come before due to the human element”, Proton said in August of its program “to better protect users who need the most security.”
That human element defines Proton Sentinel. Other than “advanced protection that will be more likely to detect and challenge suspicious events such as login attempts”, the program will escalate “suspicious events” “24/7 to security analysts who will review the assessments made by our automated systems” and escalate security support tickets to “trained security specialists” that are “experts at detecting infiltration and account takeover attempts.”
The 1200-word announcement omits that Proton does not make the most effective of account takeover prevention technology available to its customers. Instead of deploying a method that has been found to prevent 100 per cent of targeted attacks while costing a user US$50 in two security keys, Proton is charging US$120 per annum for access to a program which it does not claim carries a comparable success rate and which it would have had no basis to do so.
A 2019 joint study by Google and New York University found security keys, a second factor authentication device physically required at the point of login, prevented 100 per cent of account takeovers. Google has made security key two-factor authentication available since October 2014. Proton did so in October 2022, eight years later. While these are comparable authentication standards, they are not the same implementation. Again, Proton’s security announcement omitted its limitations, a lack of disclosure that disadvantages the risk awareness and therefore security of its users. Proton users cannot enable security key two-factor authentication, the U2F protocol, without also enabling on-device authentication, a weaker method type that Google and New York University found prevented as low as 76 per cent of targeted attacks. An attacker does not need to overcome the strongest defence to compromise a system; it needs to overcome the weakest.
Although that study did not assess how TOTP-OATH, the exact authentication standard Proton requires U2F users to also enable, performed against targeted attacks, it did find TOTP-OATH only prevented 94 per cent of non-targeted phishing attacks. The basic statistical application of these findings is that at least 6 per 100 and as many as 24 per 100 account takeover attacks against a security key-using Proton customer will succeed. Google, which does not necessitate users simultaneously enable a weaker authentication standard when seeking the highest, makes the full protection available to its users.
Proton offers security and privacy advantages that Google and its other competitors do not. Making end-to-end encrypted email more accessible is one. Security from unauthorised account access is not. It may be that Proton intends to make U2F-only authentication available once it develops its apps for compatibility, which it committed to in October 2022: “we’ll be adding support for security keys across all Proton platforms, including our desktop and mobile apps.” That it has released the labour-intensive, human-discretion Proton Sentinel a year later and ahead of the scalable, less expensive and most effective U2F-only authentication method is unexplained.
For Proton Sentinel to avert an account takeover, Proton must first electronically detect a suspicious event then a Proton employee must review and agree with that assessment. Proton has confirmed it may only intervene after the malicious login succeeds, meaning any Proton intervention may only mitigate the impact of the unauthorised access instead of preventing it: Proton “can also minimize damage by locking the account even after attacker gets in”. How Proton expects to successfully distinguish a sophisticated actor perpetrating a targeted attack from the legitimate account holder is unexplained. Malicious logins can be made on a similar device from an IP address in a similar location. Proton has released no data on its rates of false negatives despite having operated a Sentinel-like program on an invitation-basis before releasing Proton Sentinel.
Making humans the arbiters of whether an attempted login is legitimate or not exposes Proton to new malicious insider risk for which it has no publicly stated mitigations. Nor has Proton publicly addressed the jurisdictional and geopolitical risk arising from having “a global team in multiple time zones, working in shifts” appointed to this operation. Proton has not explained its decision to repose responsibility for admitting or rejecting admission to at-risk accounts in a human workforce given an alternative of a decentralised, technical solution with a demonstrated 100 per cent success rate. That Proton considers this offering “surpasses everything that has come before” poses cause to question whether the internal accounts of Proton employees tasked with performing these security judgments are protected by the same inferior methods, exposing them to account takeover risk.
Proton lists its high-risk users to include “journalists from the largest publications, governments of several countries, leaders of international peace organisations, heads of major religions, and members of parliaments.” Continuity and stability of access to their own accounts may be similarly or as important to these users as preventing unauthorised account access. These users may also be the most likely of Proton’s user base to generate “suspicious events”, such as signing in while traveling internationally, using VPNs or other anonymisers to encrypt their internet traffic and alternating between devices. “Sentinel relies upon more signals, such as device types”, Proton explained. Proton said it had experienced no false positives in its precursor, by-invitation account protection program.
Proton has not explained how it would securely authenticate and communicate with an affected user if it detects a “suspicious event” and blocks access to their Proton account. All Proton has confirmed is that it will leverage the weakest of recovery challenge methods to make these security assessments for its at-risk users: it will “leverage things like your recovery phone number or email to allow threat escalation or assessment on a case by case basis”. Google and New York University found SMS authentication prevented only 76 per cent of targeted attacks. Yet Proton has since told its users that Sentinel “provides a level of protection that greatly exceeds that which is possible via automated systems alone.”
That Proton announced U2F security key compatibility without warning users its protective effect risked being wholly negated by continued enforcement of TOTP-OATH registration does not serve the interests of its security-concerned user base, particularly those at risk of targeted attacks. Nor does announcing Project Sentinel as a high-risk security program that “surpasses everything that has come before” while omitting the U2F-only alternative with a demonstrated 100 per cent success rate, an at least 40 per cent reduced cost and without the many unaddressed risks of centralised human security assessments.
Cyber at-risk people may also be physically at-risk people. Activists working in hostile jurisdictions, journalists investigating human rights abuses and politicians opposed by foreign state-backed actors are among Proton’s user base. While Proton is for-profit and not expected to advertise its deficiencies, it is unbecoming of a responsible cyber provider founded with the express mission of privacy and internet freedom to neglect to inform its at-risk users of the live vulnerabilities in its “high-security” offering.
15 September 2023
Charges
The Australian Federal Police charged a person with a foreign interference offence on 14 April 2023.
Indeed, this case marks the second occasion someone has been charged with foreign interference under the National Security Legislation Amendment (Espionage and Foreign Interference) Act 2018.
However, that 2018 amendment introduced three foreign interference offences (s92.2-4): intentional foreign interference, reckless foreign interference, and preparing for a foreign interference offence. The first person charged was charged with the preparatory offence; the second, this person, with the reckless offence. To date, no person has been charged with intentional foreign interference.
There is no indication Friday’s charged person is accused of handling classified information. The AFP has not charged them with reckless or strict liability espionage offences related to sharing classified information with a foreign principal ((s91.1(2), 91.2(2), 91.3(1)).
Unlike those espionage offences, the chosen reckless foreign interference offence (s92.3(1)) does not require dealing with classified information nor information of any kind. The offence depends broadly on “conduct”.
The AFP chose the s92.3(1) reckless foreign interference offence that has a larger evidentiary burden and a lesser 15 year maximum imprisonment term, instead of a reckless espionage offence that covers making unclassified national security information available to a foreign principal (s91.1(2)).
The AFP alleges the person “compiled a number of reports” and “accepted payment” after being asked to “obtain information about Australian defence, economic and national security arrangements.”
The s91.1(2) espionage offence can expressly cover sharing “information or an article” that “concerns Australia’s national security ((b)(ii)) but is distinct from information that “has a security classification”. Unlike reckless foreign interference, it requires proving recklessness as to (c)(i)-(ii).
Regarding the AFP’s “economic” “arrangements” allegation, the s91.2(2) reckless espionage offence can cover sharing any “information or article” with a foreign principal. Again, this offence requires (b).
That the Commonwealth Director of Public Prosecutions is proceeding with reckless foreign interference instead of either or both of these reckless espionage offences that expressly relate to sharing information with a foreign principal, means:
They will not allege Friday’s accused was reckless as to whether their conduct will advantage the national security of that foreign country; and
They may not allege Friday’s accused was reckless as to whether their conduct will prejudice Australia’s national security.
The CDPP may instead allege Friday’s accused was reckless as to whether their conduct would support intelligence activities of a foreign principal (iii).
This would mean Friday’s accused would not have been charged without the Turnbull Government’s 2018 amendment. Before this, only alleged conduct that prejudices Australia’s national security or advantages the national security of a foreign country, whether intentionally or reckless as to whether, was in scope.
If Friday’s accused is committed to trial and then found guilty for conduct including the provision of open source information, it may be a precedent that human resource matters: that collecting and compiling, even if that information is open source, can render “support”.
The accused is entitled to the presumption of innocence.
Originally posted on X on 21 April 2023.
Defences
Alexander Csergo, charged with reckless foreign interference and the second person charged with a foreign interference offence in Australia, has released a letter, as reported by the Sydney Morning Herald. It contains claims that may give rise to three defence strategies, all of which, if the Attorney-General consents to his prosecution pursuant to section s93.1 of the Criminal Code Act 1995 (Cth), may for the first time contribute to determining what constitutes foreign interference at Australian common law. This note is not informed by the prosecution brief; it is limited to some of the select details that have been publicly released by the Commonwealth Director of Public Prosecutions and the accused.
The reckless foreign interference offence found at section 92.3(1) of the Code is not a strict liability offence. The prosecution must prove mens rea. ““The reports consist of over 20 interviews, that includes individuals from industry and ex-government employees – what is known to [Australian authorities] is that none of these interviews had taken place,” he wrote”, the Sydney Morning Herald reported. That the accused claims he produced false information in his alleged remunerated work for people whom the CDPP alleges were intelligence officers of a foreign state, while he was physically present in that foreign jurisdiction and subject to its laws, may give rise to reasonable doubt of his mens rea. If the accused did indeed produce reports containing false information, he may have been at risk of adverse action for what may be viewed by that foreign intelligence agency as fraudulent conduct or more seriously, counter-intelligence efforts undertaken in the interests of Australia, arguably ironic given it is Australia that is prosecuting him. The prosecution would have to prove beyond reasonable doubt that the accused did have mens rea despite assuming this personal risk, at a time when two Australian nationals had been imprisoned for years by the People’s Republic of China for alleged national security offences.
Recklessness at section 5.4 of the Code requires an alleged offender to be aware of a substantial and unjustifiable risk associated with their actions. The prosecution would have to similarly prove beyond reasonable doubt that the accused was aware of a substantial and unjustifiable risk of committing foreign interference by allegedly providing reports containing false information to a foreign intelligence agency. More specifically, that he was aware of a substantial and unjustifiable risk of supporting foreign intelligence activities (ss92.3(1)(iii)) or prejudicing Australia’s national security (ss92.3(1)(iv)) by allegedly contributing reports containing false information about Australia to the intelligence collection efforts of a foreign intelligence agency. To what extent all of the alleged reports and their contents were false has not been explicitly publicly addressed by the accused. Given the accused is charged with one offence instead of separate offences for each or some of his alleged deliverables, a successful conviction may depend on only one proven instance of alleged reckless foreign interference of the many alleged instances of conduct.
Although more intricate, the Court may ultimately consider whether the alleged reports could render support to that foreign intelligence agency despite their claimed lack of reliability. That the alleged reports contained false information does not necessarily render them devoid of all value. The statutory offence (s92.3(1)) does not prescribe any minimum level of support necessary to constitute an offence; even the conferral of marginal advantage may constitute supporting foreign intelligence activities. Whether the CDPP, with evidence from the Australian Federal Police and Australian Security Intelligence Organisation, would seek to prove beyond reasonable doubt that the alleged reports could support foreign intelligence activities may depend on a willingness to reveal their own intelligence of the capability of that foreign intelligence agency to have separated fiction from fact, and its capacity to have unilaterally collected the facts in the absence of the alleged contributions by the accused.
To prove an offence of reckless foreign interference, the CDPP must prove one of three characteristics of the conduct. Given the alleged offending, it is likely the CDPP would seek to prove that any part of his conduct was covert or involved deception (ss92.3(1)(d)(i)). “Csergo disputes any covert activity or deception took place – even saying it is not unusual to meet in private rooms in Chinese restaurants”, the Sydney Morning Herald reported. “When Csergo’s home was raided, officers allege they found a “shopping list” that instructed Csergo to collect information on Australia’s policy and strategy for China”, it continued. Whether allegedly meeting in private rooms or allegedly retaining the “shopping list” upon his return to Australia constitutes covert conduct may also be a decisive issue in reaching a verdict in a trial.
“Csergo’s letter says he was “suspicious and paranoid” when first approached by Ken and Evelyn in Shanghai and asked to write reports about Australian lithium, and later defence capabilities”, the Sydney Morning Herald reported. That he was approached by individuals who the CDPP allege are intelligence officers while he was physically present in their jurisdiction may give rise to the statutory defence of duress. Unlike the common law defence which is limited to threats of death or serious injury, the duress defence found at section 10.2(1) of the Code does not limit the nature of the threatened harm. The applicability of this defence may, however, be complicated by whether the accused initially freely allegedly entered into his association with “Ken” and “Evelyn” to provide strategically or politically valuable information to them, given the defence does not apply “if the threat is made by or on behalf of a person with whom the person under duress is voluntarily associating for the purpose of carrying out conduct of the kind actually carried out” (ss10.2(3)).
""[Csergo] took precautions… to ensure that no information, that was confidential or classified, was provided and that he would not engage anyone that had access to such material – and he never did,” Csergo wrote in his letter of his dealings with Ken and Evelyn”, the Sydney Morning Herald reported. Although this may reduce the objective seriousness of the alleged offending, it may be immaterial to proving an offence of reckless foreign interference. As written in a prior note, there is no indication the defendant is accused of handling classified information. The AFP did not charge the accused with reckless or strict liability espionage offences related to sharing classified information with a foreign principal ((ss91.1(2), 91.2(2), 91.3(1)). Nor did the AFP charge the accused with a reckless espionage offence that covers making unclassified national security information available to a foreign principal (ss91.1(2)) or a reckless espionage offence that covers sharing any “information or article” with a foreign principal (ss91.2(2)). This is despite the AFP media release alleging the accused “compiled a number of reports” and “accepted payment” after being asked to “obtain information about Australian defence, economic and national security arrangements.”
Unlike those espionage offences, the chosen reckless foreign interference offence (ss92.3(1)) does not require dealing with classified information nor information of any kind. The offence depends broadly on “conduct”. Also unlike the espionage offences, a conviction can be secured if the conduct was reckless as to whether it would support foreign intelligence activities. That the AFP instead selected the reckless foreign interference offence, which carries a lesser fifteen years maximum imprisonment term and a larger evidentiary burden, may indicate the CDPP assessed a conviction as more likely if the defendant was accused of recklessness as to supporting foreign intelligence activities instead of only the arguably more serious recklessness as to prejudicing Australia’s national security.
The accused is entitled to the presumption of innocence.
28 November 2023
On Friday the Foreign Minister told ABC Adelaide’s Nikolai Beilharz that she received advice from DFAT “they don’t have security concerns because they’re not connected to the internet, and they’re not connected to our own system”.
Both Hikvision and Dahua issue firmware updates via their websites, including to patch vulnerabilities. The Australian Cyber Security Centre issued a “critical” alert on 22 September 2021 to patch a vulnerability in Hikvision products “which could allow a cyber actor to take full control of the device.”
Essential Eight maturity levels two and three respectively require “security product” applications be patched for security vulnerabilities “within two weeks” and “within 48 hours if an exploit exists”.
“Security product” is named as distinct from “vulnerability scanner”. If, despite that, any department does not interpret “security product” to include security camera firmware, maturity levels two and three require patching “within one month” for “other applications”.
Subsection C.1(7) of Policy 10 of the Protective Security Policy Framework states entities “must implement Maturity Level two” “to meet the minimum requirements established under the PSPF maturity model”.
Guidance explicitly includes “applying patches to applications and operating systems of… network devices” (ss C1.2(17)). “Patches for security vulnerabilities come in many forms [including]… fixes that require overwriting of the firmware on network devices” (ss C.1.2(18)(h)).
If the 14 departments with a collective, at-least 919 Hikvision and Dahua cameras are compliant with the PSPF, they are applying firmware updates distributed from each vendor’s website. This necessitates connecting to the cameras, whether by wire or wireless.
If under DFAT, which has no security concerns, the cameras are “not connected to [its] own system” and “are not connected to… agency IT networks” (Q1084):
How can their footage be monitored? How do they practically function as security cameras?
How are they patched? Patching necessitates connecting. If unpatched, they’re vulnerable.
If they’re vulnerable, why doesn’t DFAT have any security concerns about the cameras?
Originally posted on X on 21 April 2023.
TikTok’s Global CSO spoke with CyberScoopNews’ Jeff Stone on 27 August 2020. This interview predates the “at the very least” September 2021 to January 2022 timeframe of BuzzFeed News’ Emily Baker-White’s report.
The China-based ByteDance employees who, at least after this interview, reportedly repeatedly accessed TikTok private user data - and who reportedly therefore were, or are, capable of accessing it - are subject to the Chinese Government’s law, are they not?
Why would the Chinese Government request data from US ByteDance if, at least after this interview, it could, or can, compel it from ByteDance in China, among whom there reportedly existed, or exists, an engineer referred to as “Master Admin” who “has access to everything”?
Private user data access “based on requests” is not equivalent to requests being the only method of access. How impactful is the “guise and controls of the U.S. security team” if they can be circumvented? Why would China-based employees request what they reportedly could, or can, already access?
“We have made clear that TikTok has never shared user data with the Chinese government”, TikTok’s US newsroom said in response to the Trump Administration’s 2020 TikTok executive order.
This is not equivalent to saying: “TikTok private user data cannot be accessed from China and is therefore not subject to the Chinese Government’s laws.” “Never shared data” is not equivalent in meaning to “data never accessed” if China-based employees, subject to those laws, have unilateral access.
Originally posted on X on 2 July 2022.
X post by Senator James Paterson.
In a 17 March 2022 report by The Guardian’s Rowena Mason, the UK Labour Party, Liberal Democrats, and Conservative chair of the Parliament's Foreign Affairs Committee Tom Tugendhat warned that the UK was lagging other nations in countering foreign interference.
The UK Parliament's Intelligence and Security Committee delivered its Russia report in July 2020. Quoting then-Home Secretary Sajid Javid, who said “we don’t have all the powers yet" to counter the threats, the Committee made findings and recommendations.
"The current legislation enabling action against foreign spies is acknowledged to be weak... crucially, it is not illegal to be a foreign agent in this country", it noted. Javid said the Official Secrets Act had become “completely out of date".
MI5 Director-General Andrew Parker observed that "there are things that compellingly we must investigate, everybody would expect us to address, where there isn’t actually an obvious criminal offence because of the changing shape of the threat... this doesn't make sense."
Parker observed "...today it is not an offence in any sense to be a covert agent of the Russian Intelligence Services in the UK - just to be that, to be in covert contact, to be pursuing a brief – unless you acquire damaging secrets and give them to your masters."
To enable "MI5 to defend the UK against agents of a hostile foreign power such as Russia", the Committee recommended reforms including a foreign influence scheme modelled on the U.S. Foreign Agents Registration Act.
The Committee also affirmed a February 2019 finding from the Digital, Culture, Media and Sport Committee's Disinformation and 'Fake News' inquiry that “the UK is clearly vulnerable to covert digital influence campaigns."
The Intelligence and Security committee questioned "whether the Electoral Commission has sufficient powers to ensure the security of democratic processes where hostile state threats are involved; if it is to tackle foreign interference, then it must be given the necessary legislative powers."
Following the recommendations, the Johnson Government's May 2021 Queen's Speech committed to a Counter-State Threats Bill that would "[make] the UK a harder environment for states to conduct hostile activity in and increasing the cost to them of doing so."
The Bill would also "[ensure] the security services and law enforcement agencies have the necessary powers to tackle future threats and the evolving tactics carried out by hostile states and actors."
Given the Official Secrets Act does not "address the modern threats posed by hostile activity" from states, the proposed bill would modernise both the OSA and security service powers, create a Foreign Influence Registration Scheme, and consider criminalising foreign interference.
The UK Home Office led a public consultation on the proposed national security bill from May to July 2021. The Johnson Government is yet to respond.
Referencing the proposed Foreign Influence Registration Scheme, Conservative Foreign Affairs Committee chair Tom Tugendhat said: "This is an essential piece of legislation for keeping our country safe... We need to protect our interests."
Shadow Home Secretary Yvette Cooper said: "For too long our democracy has been open to foreign interference, and the government has failed to act. Labour has been calling for foreign agent registration law to be brought forward for some time..."
These calls to action come two months after MI5 alerted the UK Parliament in January 2021 to an actor allegedly “knowingly engaged in political interference activities on behalf of the United Front Work Department (UFWD) of the Chinese Communist Party”.
Originally posted on X on 17 March 2022.
Australian Financial Review: Australia warned over online electoral misinformation; ABC News: Canadian politician warns Australian voters could be vulnerable to foreign interference; 2GB: Election influenced by propaganda & foreign interference
Part one
Approaching the federal election, the Australian Electoral Commission has moved to mitigate the threat of electoral misinformation. How resilient is Australia against both electoral misinformation and political disinformation?
Does the AEC, directly or from the National Intelligence Community-supported Election Integrity Assurance Taskforce, have adequate linguist resourcing to combat in-language misinformation? How many would be adequate? Has the AEC predicted misinformation volumes? Would more linguists be available on-demand?
On how many non-U.S. social media services, like WeChat, is the AEC present? How many Australian users and communities can it reach? What proportion of election-related content does the AEC expect to have visibility of? Can electors swiftly report misinformation to the AEC?
From how many social media services, including non-U.S. services like WeChat, has the AEC sought and received assurance that they would promptly remove electoral misinformation? Have any been unwilling or otherwise non-responsive? If so, are there alternative strategies in place?
What is the scope of misinformation that the AEC would be willing to act on? Would the AEC be strictly confined to electoral misinformation? Will any agency be live monitoring for political disinformation from foreign state actors? If discovered, what can and would any agency do?
The Commonwealth Electoral Act 1918 does not grant the AEC any power to regulate the political contents of any electoral communication. The constitutional implied freedom of political communication may limit government efforts to counter political disinformation.
Are Australian election candidates at risk of political disinformation as the 2022 federal election approaches? There is recent precedent. In Canada, disinformation about a counter-foreign interference legislative proposal was published on WeChat to Chinese diaspora communities.
Then-Conservative-MP Kenny Chiu introduced the Foreign Influence Registry Act (Bill C-282) in April 2021. Narrower than Australia’s FITS, it would have required individuals acting on behalf of a foreign principal to register influence activities directed at public office holders.
According to the the Atlantic Council’s DFRLab’s Kenton Thibaut, a WeChat account named “Toronto consulate”, which circulated daily news briefings to Chinese Canadian diaspora communities, re-published an article about the bill ten days before the September Canadian federal election.
The Chinese diaspora would be controlled, monitored, and have their freedom of speech curtailed, the article claimed. Kenny Chiu denied these claims. The article’s title urged readers to spread the information. If Chiu was re-elected, the bill may become law, it noted.
The WeChat account was operated by an employee of HuayiNet, a translation company that has PRC government and government-affiliated entities among its clients, including consulates in Canada and the U.S.
Because of the relationship the “Toronto consulate” WeChat account owner had with the actual PRC consulate in Toronto, the account and its post appeared alongside a “consulate certified” WeChat label.
According to the 2016 census, the population of Chiu’s riding Steveston-Richmond East has a 48.7% Chinese ethnic origin. There was an 8.2% swing against Chiu. Chiu was not re-elected. Causation is not necessarily implied.
If suspected political disinformation emerged in the 2022 federal election campaign, and it was suspected to constitute foreign interference, which department or agency would respond and what would it do? The Canberra Times’ Sarah Basford-Canales has explored this issue.
Attributing political disinformation to a foreign state actor or to an entity acting on behalf of a foreign principal is what distinguishes foreign interference from other political communications. What evidentiary threshold should apply in these judgements?
Disinformation timing and government resourcing may mean investigations, if any, reach no findings until after the election. Even if found to be disinformation, there may not be adequate evidence for foreign attribution. Existing laws may not provide for content removal.
While the National Security Legislation Amendment (Espionage and Foreign Interference) Act 2018 (Cth) criminalised foreign interference, a conviction for election-related contraventions may not reverse nor mitigate its time-sensitive impact.
How can Australia increase its resilience against disinformation foreign interference? These example solutions are all imperfect and pose their own risk.
Should the federal government launch a public foreign interference awareness-raising and deterrence campaign that references the new extended geographical jurisdiction criminal offences and their penalties in division 92 of the Criminal Code Act 1995?
Should political parties enter into mutual-assistance agreements to publicly repudiate disinformation affecting any candidate? This may reduce any partisan advantage from disinformation and in doing so disincentivise its perpetration.
Like the AEC has temporary election employment, should Home Affairs temporarily employ linguistically-diverse open source intelligence analysts from Australia and across the world for the duration of an election campaign and election day to monitor social media services? They could refer any disinformation to political parties in real-time for time-sensitive review and if agreed to be disinformation, for collective repudiation. The suspected disinformation could also be referred to the AFP, ASIO, ASD, and others, for further investigation.
To avert the risk of any perceived or alleged partisanship or politicisation, the intelligence unit could be configured as a statutory body within the Home Affairs portfolio, independent of any operational involvement or oversight from the Department.
Originally posted on X on 6 February 2022.
Part two
Ahead of the 2022 federal election, how resilient is Australia against both electoral misinformation and political disinformation? The Australian Electoral Commission has since elaborated on its capabilities and limitations.
The AEC social media monitoring team "do not have the capacity to monitor social media conversations in languages other than English," reported the Sydney Morning Herald’s Caitlin Fitzsimmons.
The AEC is a member of the Electoral Integrity Assurance Taskforce, which has "foreign interference or disinformation" as part of its mandate and is "supported, as needed" by the NIC. Has the AEC sought any non-English language support? Will it? Would a request be accepted?
In the 2016 Census, 4.9 million reported speaking a language other than English at home, 820,000 of which reported they spoke English “not well” or “not at all”. How will English-only electoral misinformation countering protect a linguistically diverse Australia?
The AEC has a misinformation monitoring presence on AEC TV on YouTube, Facebook, Twitter, LinkedIn, TikTok, and Reddit, reported the Sydney Morning Herald. On what criteria were these prioritised? Mediaweek says TikTok has 1.1 million active monthly Australian users but WeChat has an estimated 2.9 million.
Will the AEC or EIAT have non-English capability? Will they monitor WeChat? Only electoral misinformation is in the AEC's current scope. Misinformation about candidates or their policies is not. Will they monitor political disinformation to defend against foreign interference?
There is evidence from Five Eyes partner Canada of political disinformation circulated in Mandarin to ethnic Chinese diaspora communities on WeChat. All three elements, namely political disinformation, Mandarin, and WeChat, are beyond the scope of the AEC's current defences.
Then-Canadian MP Kenny Chiu introduced a Foreign Influence Transparency Scheme-like bill. An article in Mandarin falsely characterising it as repressive of Chinese diaspora communities was circulated on WeChat ten days before the 2021 election by a CCP-serving company's employee's consulate-certified account.
Interviewed by the Australian Financial Review’s Tom McIlroy, Kenny Chiu, who was not re-elected, warns Australia of his experience. Would a foreign interference actor be emboldened by this precedent in a liberal democracy? Is the risk greater now than it was then?
If an act of foreign interference was executed at the federal election by way of a political disinformation campaign against a candidate, would the AEC, EIAT, NIC, or any other entity have the capability to effectively act in a time-sensitive way? Would it? What would it do? How?
I note the Australian Electoral Commission's proactive efforts to correct electoral misinformation online. I note its transparency regarding its capability and limitations, without which this level of scrutiny may not be as readily feasible. The AEC is due to appear at Senate estimates at 1930 today.
Originally posted on Twitter on 15 February 2022.
Entangled Duties: Law firms operating in Australia with PRC links by Internet2.0.
ABC News: Beijing-owned businesses using Australian law firms to advise on takeovers of critical projects.