The Promise and Challenges of New Actors and New Technologies in International Justice

This article addresses the role of new technologies in the international justice and accountability landscape, drawing from research we conducted into new United Nations (UN) accountability mechanisms that have the explicit mandate to collect, collate, analyse and preserve evidence of international crimes according to criminal justice standards. The article is divided in four parts. First, we contextualize our research by discussing some of our findings and situating them against what we define as a ‘third wave’ of institutional developments in international justice, prompted by an ‘accountability-turn’ affecting civil The research for this article was conducted in connection with our project ‘Anchoring Accountability for Mass Atrocities: Providing the Support Necessary to Fulfil International Investigative Mandates’ at the University of Oxford. The project is the result of a partnership between the Oxford Programme on International Peace and Security, the International Bar Association War Crimes Committee, and the Simon Skjodt Center for the Prevention of Genocide at the US Holocaust Memorial Museum, Washington, DC. The authors are grateful to the editors of the journal and symposium for the opportunity to publish our research, and in particular to Alexa Koenig and Urmila Pullat for their input. We are also grateful to Lindsay Freeman for her insight; to our team members Ambassador Stephen J. Rapp and Sareta Ashraph for their contributions and data gathering over the summer; to all members of our project’s Advisory Group

society groups and UN mandates. Secondly, we discuss -using real-world examplesboth the opportunities and challenges arising from the use of digital and new documentation technologies in the field. Thirdly, the article pays particular attention to the role of UN mandates affected by the 'accountability-turn'; our research reveals such UN mandates now often sit at the heart of the 'life cycle' of information and evidence collected for justice and accountability purposes. In this section of the article, we also briefly discuss issues relating to third party control of information, in particular by social media companies. Finally, we discuss the need (and welcome initiative) to develop better international guidance and best practices for actors across the board in order to maximize the effective use of new technologies and digital evidence in international justice and accountability processes.

A Changing Landscape in International Justice
The past 10 years have been characterized by important shifts in the international justice and accountability landscape. 1 These include an increased exercise of universal or extraterritorial jurisdiction by domestic authorities; increased involvement of international courts and tribunals around issues of state responsibility for atrocities; and the emergence of new actors -including civil society groups and mandates of the United Nations (UN) -seeking to contribute directly to international justice and accountability processes, including criminal accountability. 2 We have termed this re-orientation of stakeholders an 'accountability turn'. 3 This trend has culminated, since 2016, in a 'third wave' of institutional developments within the international justice field, 4 namely, the establishment of a 'novel generation' of UN accountability mechanisms, which have the explicit mandate to collect, collate, analyse and preserve evidence of international crimes -including in Syria, Myanmar and by the Islamic State of Iraq and Syria (ISIS) in Iraq -according to criminal justice standards, and to make this evidence available for domestic or international prosecutions. 5 Though significant differences in mandate, jurisdiction, resources and operational realities distinguish these investigative mechanisms from one another, they may be conceptualized as 'quasi-offices-of-an-international-prosecutor-at-large', albeit not attached to a specific court. 6 To understand the challenges and opportunities offered by these changing dynamics in the justice and accountability panorama, our team at the University of Oxford has been gathering significant data and research over the past few months. 7 This includes interviews of 57 personnel working across the accountability ecosystem, from members of evidence-gathering organizations to national and international investigators and prosecutors, 8 and an anonymous survey of 103 UN staff members working with mandates affected by the 'accountability-turn'. 9 What emerges is a fascinating picture. 10 Data from our interviews indicate that as competent legal systems gain willingness and confidence to pursue cases against individuals under their jurisdiction, and as other accountability mechanisms are triggered to uphold state responsibility, UN mechanisms can perform important support functions, such as providing contextual analyses with regard to broader political and case circumstances; gathering both crime-base and linkage evidence (i.e. evidence that connects individuals with crimes); helping to collect, verify and collate information; and closing evidential/information gaps. 11 Prosecutorial authorities also recognize civil society documentation as an essential to their ability to investigate and try cases, especially where they cannot access crime scenes themselves. The increased 'professionalisation' of documentation efforts from civil society actors is also seen favourably by those ultimately seeking to 6 D'Alessandra, supra note 3. 7 See Oxford Institute for Ethics, Law and Armed Conflict, 'Anchoring Accountability for Mass Atrocities', 2020, available online at https://www.elac.ox.ac.uk/moving-fact-finding-case-build ing (visited 27 November 2020). 8 We interviewed investigating and prosecuting authorities of all 12 domestic jurisdictions most active in universal jurisdiction cases of core international crimes, as well as investigators and prosecutors at the ICC, and 34 of the leading civil society groups documenting international crimes in Africa, the Middle East, Asia and Europe. 9 S. Ashraph, F. D'Alessandra and S. Rapp, 'Structural Challenges Confronted by UN Accountability Mandates: Perspectives from Current and Former Staff', Opinio Juris, 14 October 2020, available online at http://opiniojuris.org/2020/10/14/structural-challenges-confronted-by-un-accountabilitymandates-perspectives-from-current-and-former-staff-part-i/; http://opiniojuris.org/2020/10/14/ structural-challenges-confronted-by-un-accountability-mandates-perspectives-from-current-and-for mer-staff-part-ii/; http://opiniojuris.org/2020/10/14/structural-challenges-confronted-by-un-ac countability-mandates-perspectives-from-current-and-former-staff-part-iii/ (visited 28 November 2020). 10 For a discussion of preliminary findings, see F. D'Alessandra et al., 'Anchoring Accountability for Mass Atrocities: Providing the Support Necessary to Fulfil International Investigative Mandates', Opinio Juris, 18 September 2020, available online at http://opiniojuris.org/2020/09/18/anchor ing-accountability-for-mass-atrocities-providing-the-support-necessary-to-fulfil-international-investi gative-mandates/; http://opiniojuris.org/2020/09/18/anchoring-accountability-for-mass-atrocitiesproviding-the-permanent-support-necessary-to-fulfil-international-investigative-mandates-part-ii/; http://opiniojuris.org/2020/09/19/anchoring-accountability-for-mass-atrocities-providing-the-per manent-support-necessary-to-fulfil-international-investigative-mandates-part-iii/ (visited 28 November 2020). 11 Rapp, D'Alessandra and Sutherland, Part iii, ibid. leverage digital information in judicial proceedings. 12 Indeed, our data show that many actors increasingly think of international justice as an 'evolving landscape', a complex system of information where streams of data originate with civil society groups, are processed in some form by UN mandates, and then arrive with judicial and non-judicial authorities. 13 Interestingly, civil society actors consistently rated 'criminal accountability' as either their 'top priority' or a 'very high priority' for their documentation efforts, and often looked to UN mandates to provide qualitative guidance in terms of what information is most relevant for these purposes, although they also valued capacity-building initiatives aimed at supporting the long-haul fight against impunity. 14 Of course, challenges persist. These range from discrepancies in operational and methodological approaches, to the quality and quantity of information collected. These challenges are relevant irrespective of which form of accountability is pursued. In fact, the true utility of information collected might not be apparent at the time of gathering. However, consensus is emerging that actors should seek to uphold the highest methodological and procedural standards to ensure that, if information might ever become relevant to criminal proceedings, its probative value will not be lost or compromised. 15 Our data also showed clear consensus that, as more actors and mandates operate within or around the same area, the provision of better coordination and guidance becomes crucial. Indeed, the UN Office of the High Commissioner for Human Rights (OHCHR) has recognized the crucial role that UN mandates can play in supporting accountability purposes, and has already begun to build an investigations support unit. 16 Most fascinatingly, these developments have coincided with a transformation of the information landscape that has the potential to revolutionize the field. The Internet has dramatically altered conditions of speech, becoming the latest 12  medium for the recording, publication, and indeed commission, of crimes. 17 Even when not purposively recording our actions, the digital footprint we leave behind by virtue of the technology we carry around, from our phones to our digital watches, is inescapable. In addition, intelligence sources and capabilities once the exclusive tool of states, are now publicly available to civil society groups and individuals on the frontlines, providing the opportunity to capture or prove the commission of crimes extemporaneously, even without stepping foot into a crime scene. Based on our data, four 'new technologies' are chief candidates to propel the international justice field into its new frontiers 18 : (i) geospatial intelligence and remote sensing (GEOINT); 19 (ii) online open source intelligence (OSINT); 20 (iii) financial intelligence (FININT); 21 and (iv) documentation technologies, ranging from simple photo cameras to specialized software -such as the groundbreaking 'eyeWitness to Atrocities' 22 application -that allows for the recording of photos and videos and packages those items with the relevant metadata needed to demonstrate their authenticity in court. Of course, the more tools become available to those seeking to document violence, the more guidance on how these tools can be leveraged ethically and lawfully becomes necessary. however, is only just starting. From the collection of explosive information challenging official narratives, to the verification and corroboration of information and evidence, to the co-opting of new technologies for preventive purposes, the opportunities presented by closed and open-source digital material are worth exploring in detail.

Understanding the Promise and Potential of New Technologies in the International Justice and Accountability Ecosystem
The advantages of new technologies, digital documentary and investigative methods, can be significant. 24 Collecting and transmitting evidence digitally have been lauded as having the potential to 'democratise the process of human rights fact-finding' by opening it up to 'ordinary people.' 25 This is because such methods are open to anyone with the right equipment -a smart-phone, a camera, a Facebook account or similar -and access to the Internet. Opensource investigation has the power to overcome or at least mitigate biases identified in 'traditional' investigative methods, for example, selection bias in witness interviews, or confirmation bias in investigative decisions. Authentic photographic evidence contains more visual details more accurately than eyewitness testimony ever could. In addition, because digital technologies can theoretically be used anywhere, they permit meaningful monitoring, documentation and investigation of possible atrocities by investigators remotely even for so-called 'black hole' environments, where information is deliberately hidden by local authorities or otherwise scarce.
This information cannot, of course, entirely substitute more traditional forms of evidence. However, as discussed below, it can act as a 'force multiplier' for other evidence. For this reason, it can and has infinitely expanded its potential uses, including but not limited to judicial and non-judicial accountability, including criminal and civil proceedings, other transitional justice strategies, truth and reconciliation efforts, memorialization and restorative justice processes.
GEOINT documentary technologies -including satellite imagery, radio, radar and other forms of remote sensing capacity, such as commercial unmanned aerial vehicles -can be used alongside OSINT methodologies to 24 27 Contradicting Chinese government claims, GEOINT has indicated the rapid recent development of detention camps holding an estimated 1-1.5 million Uyghurs and other Turkic minorities in Xinjiang. 28 Similarly, satellite imagery, corroborated by witness affidavits, has not only unveiled the operation and reporting structures of special political prisoners' camps (kwanli'so), north of Pyongyang, where 80,000-130,000 people are currently being detained, but also established the exact location of camps 14, 15, 16, 18, 22 and 25, the very existence of which the North Korean government has vehemently denied. 29 While measures can be taken to 'fool' satellites, and simple 'birds eye' views may not adequately show destruction or erosion of building complexes, the prevalence, independence and consistency of satellite oversight -no longer the preserve of powerful state authorities -renders it more difficult to sustain denials of wrongdoing. New technologies can also be used for early-warning and preventive purposes. 30 FININT, GEOINT and OSINT (including social network and big data analysis) can assist with tracking the movement of individuals or groups (including militias and refugees), or even the 'mood' of specific groups 31sometimes being able to predict with amazing precision the outbreak and location of identity-based protests or other atrocity risk factors. 32 FININT can also assist with criminal deterrence (through the freezing of assets and other financial sanctions, the issuing of travel bans and the exposure of money laundering 33 ), thus diminishing perpetrators' agency and structural opportunities. Finally, by bringing hidden and underrepresented voices to the fore, the 'democratising' effect of technology 34 can overturn state narratives, allow greater oversight and transparency, and significantly contribute to shaping public policy. Within international justice, these technologies hold enormous potential to ensure public, political and criminal accountability for atrocities. In the following sub-sections, we discuss some illustrative and practical examples.

A. Ensuring Public Transparency and Accountability
One example of the use of new technologies and digital documentation to promote public transparency and accountability is the work of the Londonbased watchdog 'Airwars', which monitors and documents allegations of civilian harm caused by international military airpower in Syria, Iraq, Libya, Yemen and Somalia. Airwars monitors open source material including relevant media and social media, uploaded footage of international strikes, local casualty-monitoring reports and militant propaganda; by doing so, it has been able to highlight stark divergences from official casualty counts reported by state militaries, including the anti-ISIS US-led Coalition. 35 Challenges to official reports posed by Airwars and others seem to have encouraged greater military transparency, at least with coalition-powers that engage with such findings. The Pentagon, for example, reported that in 'continu[ing] to refine its practices and procedures for reviewing reports of civilian casualties, . . . it considers reports available from any source, including after-action reporting of military units, and information provided by external sources, such as NGOs, the news media, social media and individuals who were present during the operation, including military personnel and local civilians'. 36 In addition to promoting improvements in standards, the publication of these reports also vastly  and later to identify those likely responsible for the downing of the flight. 42 When the Ukrainian Security Service and the JIT publicly released telephone intercepts, independent investigators were able to assess (and confirm) their authenticity against open source material, leading to the identification of further potential suspects. 43 In a similar manner, while the Syrian government denies ever having used chemical weapons, analyses of digital open-source information and declassified intelligence allowed the French government to determine 'with a high degree of confidence' that lethal chemical attacks occurred in Douma, Syria, at the hands of the Syrian regime. 44 In reaching this conclusion, French authorities relied on reports issued by civil society organizations as well as media, authenticating and verifying these through forensic examination and against witness testimony, finding that 'the spontaneous circulation of these images across all social networks confirms that they were not video montages or recycled images'. 45 While assertions by states seeking to attribute international wrongful conduct to other states must always be handled with care, the references to civil society and open source material in both the French and Dutch/JIT reports demonstrate very clearly the value of publishing on-the-ground information in places otherwise inaccessible to international investigators. Further, the enforced methodological transparency meant that the French and JIT/ Ukrainian claims could at least themselves be subject to rigorous scrutiny.
In another example, when pursuing its case against Myanmar for violations of the Convention on the Prevention and Punishment of the Crime of Genocide at the International Court of Justice (ICJ), The Gambia relied on satellite 42    open source -has also ensured convictions in the Democratic Republic of the Congo for rebel militia of the Democratic Forces for the Liberation of Rwanda. 58 This non-comprehensive list is but one indication of the opportunities offered by new technologies with regards to the investigation of international law violations, including atrocity crimes. As others have pertinently observed, 'the future of [accountability] will [inevitably] be intertwined with the advancement of technology'. 59 For this reason, it is imperative that anyone seeking to leverage these new technologies to document atrocities -whether states, UN mandates, courts and tribunals, or civil society groups -at a minimum understands the specific security needs and vulnerabilities of this kind of information. Operational protocols and information management tools will be necessary to allow the confident handling of information new and open-source technologies can offer. 60 In the following section, we turn to discussing some of the challenges presented by this material, and how those challenges can impact the 'life cycle' of digital information and evidence in support of accountability.

Important Challenges Presented by Digital Open-Source Material
If, as mentioned above, the opportunities offered by open source and other digital materials seem boundless, it is equally important to bear in mind that these information and documentation tools also present challenges. These range from the potential widening of the digital divide; to possible biases (towards the digitally visible, or emanating from intrinsic characteristics such as the mandates or identities of information providers); to challenges to authenticity, integrity and many others. 61 The passing of this information between parties, and the need to rely on the cooperation of third parties (including, often, private sector entities designed to profit from data) -can significantly compromise the value of this information. Appropriate measures and standards are critical to ensure the integrity of information across its lifecycle, from collection to transmission, analysis, disclosure and long-term storage. The breach or failure of protocols and other safety measures can be fatal to criminal accountability efforts.

A. Collection and Sharing of Evidence
The challenges posed by new and documentary technologies begin with collection. Access to the Internet -critical to information sharing -is subject to economic and political dynamics, including prohibitive pricing and Internetshutdowns. These impediments are of very real concern to UN mandates: in September 2019, for example, the Bangladeshi government banned 3G and 4G Internet and the use of SIM cards by Rohingya refugees in Cox's Bazar, citing security concerns, restoring Internet access only on 29 August 2020. 62 This caused a gap in collection of evidence potentially crucial to the UN Independent Impartial Mechanism for Myanmar (IIMM). 63 Unequal access to Internet services can also foster biases in collection, in the same vein as collectors' agendas and mandates.
For example, in determining what is collected, civil society organizations may be influenced or constrained by funding conditions, for example, encouraging particular focus on certain violations over others. In addition, over-reliance on 'the observable' -a bias towards the visual in documentation -may neglect less visible violations such as gender-based violence, starvation, or crimes against children. Simultaneously, online misinformation and disinformation campaigns may shape understandings and narratives of situations, and thus influence witnesses' 'perceptions of events' 64 in ways that veer from the 'truth'. This means that deliberate strategies must be deployed not only to ensure that evidence-collection represents the landscape of a situation as comprehensively as possible, but that any potential for bias is addressed and accounted for. A strategic approach to investigations that pays particular attention to 'less visible' crimes, to monitoring of local reporting, and to reliable eyewitness and survivor accounts that can be accessed securely can assist in compensating for blind spots and asymmetric power balances. locating, selecting and analysing any evidential material, vigilance regarding structural and cognitive biases, including awareness and care regarding algorithm design and input, 65 and a scrupulous scepticism regarding the authenticity of digital material 66 will all contribute to ensuring that information collected and shared with accountability mechanisms maintains its independence and integrity. Verification processes designed to minimize the introduction of bias into forensic analyses are being developed by specialized investigators. 67 International accountability institutions, including UN mandates, can play a valuable role in promoting such best practices.
Incorporating digital safeguards at the evidence collection stage is key, though not always practicable. The development of apps such as 'eyeWitness' goes a long way to overcome issues of security and integrity of material, allowing the instant sharing of on-the-ground data from the most inaccessible places. Our data show, however, that barriers of digital access, digital literacy and trust persist. 68 The risks of taking and uploading such material may be very high. For this reason, safety and security protocolsand the 'do no harm' principle -should always be of guidance. 69 In addition, much digital material is lost on damaged, stolen or seized devices, or deleted. This calls for 'contingency plans' to ensure that an authenticated copy (or original) of the material remains available. 70 For example, YouTube's community standards mean that 'violent or graphic' content will be removed; while exceptions apply to 'newsworthy' content, without back-up, it is inevitable that material potentially valuable to criminal investigators will be lost. 71 For material uploaded to social media sites, the retention and creation of metadata -including that of unpublished material -would be a simple means of safeguarding authentication ability.

General Challenges
In a 'post-truth' world, the camera often lies. Like other forms of potential evidence, digital material may be manipulated or destroyed, and authorship hidden, denied or falsified. Disinformation campaigns ('the systematic use of deliberately distorted information to manipulate an adversary's decision-making elite, or public opinion') exploit existing social divisions and biases, relying on 'unwitting agents' to spread false narratives. 72 This became evident in the Syrian conflict: despite -or perhaps because of -being one of the most heavily recorded conflicts in history. Disinformation campaigns have used faked, doctored and plagiarized material to malign and discredit various actors. 73 International mechanisms scrutinizing conflicts involving major states are also likely to be targeted with disinformation and 'false flag' campaigns to distort and impede investigative efforts. Clarifying the truth in such an environment is extremely difficult and requires well-informed understanding of the digital realm and of local dynamics, as well as sophisticated wariness incorporated into review processes and the staffing practices of monitors. 74 By carefully examining metadata (where available), and/or conducting reverse image searches, investigators may be able to detect misinformation.
One example is the case of images published by international media that purported to show Buddhist monks burning Rohingya victims. The authenticity of the images was soon debunked when it was demonstrated they actually depicted the cremation of victims of China's 2010 earthquake. 75 As noted above, however, news spreads fast on social media sites. Before fake information can be found, challenged and removed, it might be leveraged to incite hatred, or otherwise stoke the conditions for mass violence. 76 It is then critical evidence in itself. 77 In addition, 'deep fake' 78 technology presents extraordinary opportunities for disruption, falsification or fabrication of convincing 'footage' implicating or exonerating individuals, groups or militaries. As 'President Obama' (impersonated by comedian Jordan Peele) explained in a Buzzfeed video, 'we [have entered] an era in which our enemies can make anyone say anything at any point in time'. 79 Deep fake detection benefits from artificial intelligence (AI) that can forensically analyse digital material. For example, by training machine learning systems with ISIS propaganda videos until the AI could 'recognise' their distinctive features (perhaps imperceptible to human detection), machines have been taught to detect ISIS propaganda. 80 The high volumes of material required to train the system make this a feasible option for detecting falsified or fraudulent material in data-rich situations, such as Syria, but less so for situations with lower volumes of suitable 'training' material, such as the examples of 'information black holes' mentioned above. It is a pressing concern, early attempts to develop AI detection have not been wholly successful and are already outdated, 81 and the Internet's pursuit of deep fake detection tools must be monitored closely. 82 As international and domestic courts increasingly rely on digital open-source material, authentication protocols are needed that will allow actors at every stage of proceedings to ensure and assess that information's integrity and reliability. AI and other technological aides will be helpful for collecting, preserving, processing, authenticating and verifying digital material. However, algorithms and machine-learning tools will have to be designed with acute care to minimize biases, and be subjected to frequent rigorous examination to ensure they remain as objective as possible.

Criminal Justice Challenges
International legal mechanisms must accept that the online realm and the nature of modern evidence will impact their structure, staffing, processes and equipment. In this, they must follow the lead of those pioneering opensource investigations, including intelligence agencies and journalists. Unless material is authenticated and verified at the earliest stages, it will potentially be impossible to rely on it at trial. Worse, it could contaminate proceedings. Criminal procedural standards -and the ethical obligations of lawyers and criminal investigators -are higher and stricter than the obligations upon journalists, intelligence services or 'freelance' investigators. Investigative mechanisms, international and domestic courts, and other judicial authorities will be able to satisfy their mandates only if material is collected, preserved and analysed in accordance with these evidential standards. For legal actors, verification is a process of weighing probative value, not simply a binary assessment of 'true or false'. Non-digital evidence must also be proactively pursued, not least because the 'voices' of victims must not be eclipsed by over-reliance on the digital. Verification processes must still aim for the triangulation of documentary, physical and testimonial evidence. 83 This applies to accountability mechanisms across the board.
As courts adjust to digital open-source material, new norms may emerge regarding indicia of authenticity and reliability. For example, in a step hailed as a breakthrough for digital open-source investigations, the ICC relied heavily on videos posted on social media sites to issue an arrest warrant against Mahmoud al-Werfalli in August 2017. 84 In the second arrest warrant issued against him, the Pre-Trial Chamber held that 'an expert report [. . .] prepared by a renowned, independent institute' which 'concluded that there no traces of forgery or manipulation in relation to locations, weapons or persons shown in the video', and was corroborated by witness evidence, amounted to 'sufficient indicia of authenticity'. 85 Similarly, in view of the varying evidentiary thresholds of national authorities and international tribunals, 86 the recently established international investigative mechanisms for Iraq, Syria and Myanmar are inevitably being tasked with (and our data indicate often expected to) 87 assess the probative value of the material they gather. As 'core processors' of digital information, these mechanisms have begun to incorporate protocols and 'stateof-the-art' tech-infrastructure designed to support all types of accountability measures. Other mandate holders should be given the same instruments, as the work of the Myanmar FFM demonstrates, by working to the highest judicial standards, such bodies will be able to optimize the utility of their assessments and investigations, feeding directly into judicial proceedings. 88 As a potential approach to these issues, and taking instruction from opensource investigation groups, 89 we submit that one option would be to adopt the 'tiered categories of probative weight' approach to digital open-source material used for investigations. A tiered approach is more conscious of information asymmetries, digital divides and blind spots that can define some of the power structures of digital and open source material, recognizing that unverifiable sources or material may be an indication of a problem with an investigation rather than the credibility of the material. 90 Such an approach would be most helpful to national authorities, in part because it would provide a provisional assessment of probative value and would accommodate different judicial systems and rules of evidence. For example, information 'conceded' (i.e. the veracity of which is acknowledged by the belligerent/organization/individual to whom it is attributed) could be flagged as 'confirmed'. Where multiple credible sources, including, for example, biographical information, photographic and/or video material forensically authenticate the same information, this could be assigned 'high' probative value. One tier below could be 'fair', where multiple credible sources have corroborated the information, but not to the strength and standard that would assign it 'high' probative value. Finally, where information comes from a single source, this should not be disregarded necessarily, but 'kept' on the back-burner and assigned 'weak' probative value -leaving only 'discounted' materials, that is, those for which allegation or preservation is not considered credible -for the 'dustbin'. Our interviews indicate that domestic authorities would welcome such an approach.

C. Archiving and Storage
The archiving and preservation of digital and open-source material, of course, present its own challenges to integrity, and to loss protection (either physical, or of probative value). This raises challenges concerning the 'security of premises and archives', some of which are technological, operational and structural, and others which will require legislative and political solutions. 91 For example, digital material held by private companies incorporated in various countries raises issues regarding ownership, vulnerability to state access -for example, through 'search and seizure' orders -transparency, privacy and public accountability. Material sourced electronically may also be monitored, raising very real security concerns for activists. Benefitting from UN privileges and immunities, the archives of UN mandates are 'inviolable' to states interference. 92 Although, they must of course ensure that technical, physical and procedural safety measures are in place wherever this information is stored. These issues require urgent consideration, and call for a dedicated study and bespoke solutions. While exhausting these issues lies beyond what we can achieve here, we use the next section to address a few.

The Potential of Investigative Mechanisms and other UN Mandates
The legitimacy (both legal and political) of accountability measures relies on their accuracy and impartiality. This is particularly the case when criminal accountability is involved: the political, legal and reputational ramifications for trials and convictions based on false evidence can be devastating. The novel UN investigative mechanisms for Syria, Daesh and Myanmar -as well as, increasingly, other UN mandate holders -play an important role in this ecosystem of information as they centralize the exchange of data, whether the data originate from civil society sources or domestic and international authorities. For this reason, the mechanisms stand to play a crucial role in promoting the secure collection of authentic material; analysing and storing evidence; and encouraging and enabling the effective pursuit of accountability measures, including beyond the courtroom. By developing frameworks for cooperation with judicial and non-judicial mechanisms, these bodies can encourage reconciliation, or feed into reparation and restoration processes. Our interviews with evidence-providing organizations operating in unstable regions suggest an appetite for rule of law and judicial capacity building initiatives as a step towards more comprehensively combatting impunity. 93  especially with regard to individual accountability that these mandates can perform a precious -and unique -function. Mandate holders also stand to serve a crucial role as repositories entrusted with the storage of material obtained at great personal risk, and tasked with its navigation, analysis and archiving. The consistent development of protocols with civil society actors would optimize the mapping of violations and abuses, and ensure the safe and secure provision of information within their possession. At the time of writing, a great step in this direction is constituted by the Protocol of Cooperation between the International, Impartial and Independent Mechanism and Syrian Civil Society Organizations participating in the Lausanne Platform. The protocol clarifies and encourages active engagement between the IIIM and Syrian NGOs, in order to facilitate the shared goals of pursuing justice, accountability and redress for victims. 94 The 'two-way dialogue' allows all parties to focus on maximizing their contributions, whether the IIIM providing support to civil society documentation and analysis, or affording the evidence providers better understanding as to what the most valuable evidentiary material is, and how to obtain and provide it in a way that meets criminal legal standards. Protocols, tailored-Memorandums of Understandings, and collaborative relationships may go a long way in establishing and maintaining trust to overcome security concerns regarding the transmission of digital material.
Downstream, that is, facing the 'consumer' end of this flow of information, the novel mechanisms -and other UN mandate holders with the support of OHCHR -could fulfil an equally important function. The development of protocols or cooperation agreements with national and international investigation and prosecuting authorities encourages and streamlines engagement between these parties, and ensures that the information collected, received and analysed by UN mandates supports the independent work of forensic investigators and criminal prosecutors. Our data indicate that awareness of the role that can be played by UN mandates is increasing because of the performance of novel investigative mechanisms, and that national authorities believe these mechanisms can play a useful role in providing linkage evidence and authoritative contextual analyses to guide their own case-building. 95 In view of the likely volumes of digital material (whether 'originally' digital or 'hard' evidence digitally reproduced), the ability of such bodies to provide rapid, useful responses to Requests for Assistance will be reliant on sophisticated legal software, meticulous digital archiving and well-considered 'tagging'. Our data also suggest that information provided to national court proceedings by UN bodies may -by virtue of its perceived 'stamp of UN approval' -have been accorded a higher level of trust than it would otherwise had it come from a different source. 96 This, of course, carries intrinsic risks where UN mandates do not employ proper authentication and verification procedures. At the same time, if such capabilities were provided by re-purposing some of the 'state-ofthe-art' technology possessed by the novel investigative mechanisms -which our data indicates is a possibility 97 -these needs could be met with minimal effort, and the initial investment into such technical capabilities would be put to larger (and therefore more impactful) use. How this could occur is a question ripe for scrutiny.

A. Third Party Control of Digital Information and Evidence: The Thorny Question of Social Media Companies
The challenges posed to the 'lifecycle' of digital evidence and information by commercial third parties who possess such material but are not oriented towards using this material to support accountability are also significant. Indeed, the identification, retention and collection of material held by social media and other tech companies (SMTCs) is a pressing concern for accountability efforts, and must be addressed with urgency. 98 The case of Myanmar is illustrative. Despite being described as an 'information black hole' due to the Myanmar authorities' stark exclusion of independent observers from Rakhine State, digital open-source material has enabled researchers to investigate remotely: Google Earth satellite imagery has confirmed attacks on specific villages; NASA's active fire data recorded episodes of live fire as villages were razed; and images and videos uploaded to social media could be geo-located with satellite imagery and cross-corroborated with eyewitness accounts from Rohingya refugees. 99 Since its introduction in Myanmar in 2010, the Internet (predominately Facebook) became the medium for a wildfire of misinformation, hate speech and incitement to violence against the long-persecuted Rohingya minority, and is considered to have significantly contributed to the latest wave of extreme, genocidal, violence. 100 Facebook eventually announced that it had taken steps towards the 'proactive detection and removal of hate speech', removing 64,000 pieces of content in the third quarter of 2018 alone. 101 However, to many, this was too little and too late.
In addition, the deletion of such information, in particular, when detected by algorithms, has become an obstacle to accountability. Incendiary rhetoric posted online is primary evidence of obvious value to those investigating both state and individual responsibility for the apparent commission of international crimes. Its preservation is necessary for accountability, but is in the hands of private SMTCs. While reporting that it is 'coordinating' with the IIMM 'to provide relevant information, 102 Facebook remains entangled in legal wrangling with The Gambia, which seeks removed content for its ICJ case against Myanmar for failure to prevent genocide. 103 The plausibility of the capture, preservation and transmission of this material is indicated by the commitments Facebook has made to cooperate with the IIMM. 104 However, it remains to be seen how effective this is, and what improvements might be necessary.
In another example, of 7,872,684 videos removed by YouTube in July-September 2020 for violating its Community Guidelines, 7,390,963 (93%) were removed after automatic detection. It remains unclear, however, if and how this information is being preserved to support accountability related to the commission of atrocities. 105 Digital open-source material is promoted or deleted subject to the internal rules and processes designed for commercial purposes.
conservative voices), and by some European countries, including Germany and France. 107 Beyond this, however, a significant question is whether and to what extent Internet platforms can identify and preserve potentially relevant material that can serve accountability purposes. Of particular concern to accountability efforts is the algorithm-determined removal of content that contravenes 'user standards', or is illegal. 108 Material that would otherwise have been in the public domain is shielded from the view of investigators. This raises a series of urgent questions for accountability efforts and SMTCs alike: How is material of evidentiary interest and value to international criminal proceedings to be identified and retained? Should the platforms be obliged to alert law enforcement authorities upon receipt of material that crosses a defined threshold, and, if so, which? How, and with whom, should the platforms share this material? Can such material be protected by immunities or subject to state search and seizure orders? If so, subject to which jurisdictions? For how long should platforms be asked to store tranches of material that might be of evidentiary value to a possible, distant future legal process? Which legislative tools (e.g. akin to CLOUD Act Agreements, or Mutual Legal Assistance Treaties) need to be developed to ensure that international mandate holders can efficiently and comprehensively access material held by third parties while respecting data rights?
While some resources, such as the above-mentioned Berkeley Protocol, are making crucial contributions to this rapidly developing field, our findings indicate that the development of further, discrete guidance will be required as the technology and field evolves, and challenges become more apparent. As the gatekeepers to much of the digital realm, SMTCs must be part of the conversation, alongside lawyers, ethicists and lawmakers. Because they will ultimately control how their algorithms and policies work, it is imperative that they are engaged early and promptly by (and, in turn, themselves engage with) accountability actors across the board. Some efforts are already underway, 109 but more are needed. The continued development of international guidance and best practices for those operating at every stage of the evidentiary flow is crucial to optimizing the probative value of material generated and gathered by those caught up in the world's most atrocious crimes. This is, we believe, today's biggest challenge in the field of accountability and international justice.

Conclusions
As our data indicate, a 'third wave' of institutional developments in international justice -prompted by what we have termed the 'accountabilityturn' -is reflected in the shared goals and increasing symbiosis between evidence providers on the ground and accountability actors around the world. The spread of OSINT, GEOINT, FININT and documentary technologies beyond the control of state apparatuses has revolutionized how evidence and data move around information systems, forcing accountability actors to contend with new realities. Effectively harnessing this digital revolution will be the key to the success of efforts to pursue public transparency, and to holding individuals and states accountable for the world's worst atrocities.
As this article has discussed, the opportunities are extensive. Howeverfrom disinformation and false flag attacks, to verification and authentication of digital evidence, to the potential for biases in and vulnerability to manipulation, to challenges to the assessment of digital information's probative value, to secure information management and archiving -the challenges offered by new technologies and digital evidence are significant.
UN mandates increasingly sit at the heart of the 'life cycle' of digital information and evidence. As such, such institutions stand to play an important role in overcoming the many challenges that this changing landscape and ecosystem presents. Our Oxford project 'Anchoring Accountability for Mass Atrocities' aims to contribute to this discussion. This includes understanding what could be within UN mandates' capabilities (and remit) with regards to the verification and safe storage of material, and how this can be shared with other accountability actors and institutions. However, urgent challenges persist beyond these mandates. These range from the standardization of collection, verification and security processes for documenters on the ground, to third party control of information. While important steps have already been taken, additional international guidance and best practices are needed for the many actors now seeking to leverage new technology and digitally derived evidence in international justice and accountability processes. We believe that finding ways to work with tech companies, in particular, to identify and preserve information that can assist accountability efforts, is one of the greatest next challenges for this field. The challenge is ripe for scrutiny, and we welcome the opportunity to participate in these discussions, sometimes by asking hard questions -though always with the hope of supporting and improving the delivery of justice for the world's worst crimes.