Given the issue laid out above, this comment addresses whether the ICC has a reliable means to work with the OSPs, and the countries in which the OSPs are incorporated, in order to access data that has been removed from public view for violating community guidelines and, barring that, whether there are alternative measures the ICC can take to ensure the reliable and real-time collection of evidence posted online. Further evidence will first be presented that major online service providers use community guidelines to remove potentially relevant evidence posted on their sites. Following that, the discussion turns to the potential ways through which the ICC can work with the American and Irish governmental and legal systems specifically, as the main OSPs are American firms with Europe, the Middle East, and Africa (EMEA) Headquarters in Ireland, and whether those avenues allow for enough access for the ICC to retrieve any relevant evidence from an OSP server. Finally, two alternative actions that could be undertaken by the ICC to protect its interest in online data are presented and analyzed; first, potential bilateral cooperation agreements with OSPs, and second, the establishment of a new online forum under the purview of or otherwise attached to the ICC for people with evidence of atrocities to submit their evidence directly. In the end, no one potential action is sufficient to solve the issue: this comment proposes that the most effective manner through which the ICC can ensure that it has access to any relevant digital evidence is to establish its own private forum dedicated to the collection and storage of online data, maintain dialog with the necessary OSPs until such a time as the ICC program becomes viable, and pursue what national and international legal or governmental cooperation is necessary.

II. The Problem: Facebook and YouTube Removes Evidence of War Crimes

Although the ICC has yet to formally open an inquiry into the various accusations of atrocities committed during the conflict in Syria, a preliminary examination into the actions taken against the Rohingya population in Myanmar was opened in 2018 and the OTP was granted the right to investigate in 2019.6 Both conflicts present a real possibility that crimes against humanity, whether war crimes or an outright genocide charge, have occurred and the ICC has been called upon to investigate the actions of multiple parties in the Syrian conflict through the same procedural stance by which it is now investigating the conflict in Myanmar. In addition, both conflicts illustrate the potential for the use of digital evidence to corroborate the occurrence of atrocities and the problems that stem from posting such evidence on major OSPs.

A. Facebook

Despite its effective use in other human rights contexts (i.e. the Arab Spring), Facebook has established some community guidelines that have produced negative consequences for activists seeking to shed light on ongoing crimes against humanity in both Syria and Myanmar. Facebook’s community standards have specific provisions against posts depicting or containing references to violence and criminal behavior, the general exploitation of humans, and sexual exploitation specifically. Under the violence and criminal behavior standard, any posts that show or allude to organized hate, mass murder, and organized violence or criminal activity are considered a violation and are subject to removal.7 Moreover, Facebook’s guidelines against posts depicting human exploitation requires the removal of posts that show or incite the use of child soldiers8 and any posts that allude to or portray sexual exploitation.9

While those policies may be reasonable and even necessary, they have already been used to remove potentially relevant digital evidence that could be used by the ICC in ongoing or potential future investigations. The story of one Syrian man, Abdulsalam, who witnessed and recorded the bombing of a bakery in al-Bab, illustrates the issue perfectly. In January 2014, the Syrian army dropped a barrel bomb on a bakery near Abdulsalam’s mosque (the Syrian army had a pattern of using barrel bombs on bakeries in particular) in what he described as “a particularly horrific bombing.”10 He took a few pictures of the helicopter that dropped the bomb, then went directly to the scene, where he took pictures of the destruction of the bakery in which he captured images of the civilian victims of the bombing. He uploaded those images to his Facebook account promptly following the attack, believing it to be the most effective way to preserve the pictures and spread the word about the true toll of the attacks.11

Avi Asher-Schapiro, writing in the Intercept regarding Abdulsalam’s posts depicting the al-Bab bombing and its aftermath, details Facebook’s response to his posts and highlights the main issue with Facebook’s policy:

Seven months later, Abdulsalam got an automated email from Facebook notifying him that the images had been removed. Other users had complained that his photos were too gory. By the time he got the email, Abdulsalam’s other copies of the pictures were gone; his hard drive had been burned, along with his small office, when the Islamic State stormed al-Bab and Abdulsalam fled across the border to Turkey.12

In addition to the primary issue of losing direct access to Abdulsalam’s firsthand evidence, Facebook’s deletion of his pictures compounded the issues of evidence collection in Syria particularly. Specifically, the Islamic State had begun to move into Aleppo by the time that Abdulsalam took his pictures. While that sounds tangential, human rights NGOs on the ground in Syria, including Human Rights Watch (HRW), which had been documenting evidence of potential war crimes in Syria, had been pulled out prior to the January 2014 barrel bomb attack.13 One of HRW’s researchers on the ground, Ole Solvang, recognized the potential importance of Abdulsalam’s removed evidence when he said, “if there ever is a trial, this is the stuff that could become important evidence.”14

Moreover, the same situation played out in a more general sense when, starting in September 2017, Facebook began removing images posted by Rohingya Muslims regarding their treatment in Myanmar. Rohingya activists, including the Irish-based Mohammed Rafique, reported that their Facebook accounts and pages were suspended together with the removal of content they posted depicting images and videos of torture and killings in Rohingya villages.15

As evidenced by the removal of significant evidence of potential crimes against humanity in both Syria and Myanmar, Facebook’s community guidelines establish a system through which pertinent data posted online is removed from access by the general public before ICC investigators have a chance to find it. Despite amending its guidelines in 2016 to allow violent or graphic content that is “newsworthy, significant, or important to the public interest,” the reporting structure, through which users individually flag content that they feel violates Facebook guidelines which are then reviewed by Facebook employees, still allows important potential evidence to slip through the cracks. Moreover, there is every potential that individuals who have a Facebook account and the desire to protect or support individuals committing war crimes can flag evidence against those perpetrators and have it removed, thus making it far more difficult to ever find the evidence in the first place. Therefore, Facebook’s current structure makes the potential for evidence to be removed before the ICC can open investigations (as the Myanmar evidence was removed in 2017, prior to the OTP opening a preliminary examination in 2018) all too real, especially as people likely still believe, as Abdulsalam did, that Facebook is an effective tool to memorialize data.

B. YouTube

In a similar vein to Facebook, YouTube, through its parent company Google, has also established a series of policies meant to safeguard its site from violence, graphic content, and actions taken out of hatred for a specific group. YouTube’s policy specifically states that it does not allow any footage, audio, or photos of the aftermath of war, the aftermath of a terrorist attack, sexual assaults, torture, protests, riots, or corpses with significant injuries like detached limbs.16 YouTube also removes content promoting violence or hatred against groups based on a series of group identifiers (e.g. race, religion, gender).17 Both policies require the removal of any offending content.

While the issues with Facebook’s structure for enforcing its community guidelines causes issues due to its user-driven nature, YouTube’s system of using artificial intelligence to comb through videos posted on its site that violates its community guidelines against extremist or disturbing content has yielded similar negative consequences for the preservation of evidence of war crimes posted online. In 2017, the same year that Facebook was removing data and suspending accounts associated with images depicting the violence against the Rohingya in Myanmar, YouTube’s artificial intelligence system shut down the pages of 900 groups documenting atrocities in Syria. Included in those groups whose information was removed from the site were organizations focusing on analyzing images coming out of conflict zones, organizations tracking the toll of air strikes in Iraq and Syria, and local Syrian media organizations like the Idlib Media Center, which was one of the last groups producing content from rebel-controlled territory at the time.

When evidence of crimes against humanity is posted on their site, Schapiro writes:

[S]ocial media companies have over and over again chosen to ignore, and, at times, disrupt the work of human rights groups scrambling to build cases against war criminals.18

This is borne out through both the Syrian and Burmese examples, which show that, whether intentional or not, online service providers are likely going to remove evidence of the types of crimes under the purview of the ICC posted by their users because of policies put in place to combat other disturbing global trends. Moreover, OSPs are not allowed to share content data, only metadata, which only includes general account information and the location from which the evidence was posted. The content of the post is sequestered in the company’s servers and becomes all the more difficult to access once it has been removed from the platform. Given the competing forces that pressure OSPs to remove content that could otherwise be used as evidence of crimes against humanity at trial in the ICC, and given the difficulty of accessing said content once it is removed, the ICC needs to develop means of cooperating with either the OSPs or the countries in which those OSPs are incorporated to ensure that critical data is not lost in the name of community safety online.

III. Working with National Governments to Regain Removed Online Evidence, Potential and Pitfalls

In order to recover data that has been removed under the community guidelines of any OSP once it has been removed, ICC investigators would have to gain access to the provider’s servers or have that provider send them the data from the server on which it is backed up. As the ICC at current does not have any agreements with major OSPs to protect information that could be valuable, and because the major OSPs are American firms (and the US is not a state party to the Rome Statute which mandates member state compliance with ICC investigations), it may have to negotiate with the national authorities holding jurisdiction over the location of the server in order to obtain the removed data. This section outlines what that process could look like and the drawbacks to the approach generally.

A. United States

In order to understand the regulatory framework the ICC would have to navigate in order to obtain American help in retaining evidence posted online and removed under OSP community guidelines, it is necessary to examine the Stored Communications Act (SCA), which is found in Section II of the Electronic Communications Privacy Act.19 Under the SCA, which limits governmental and institutional access to stored electronic records, US domestic law enforcement may request information through an order or after the issuance of a subpoena or a warrant, depending on the type of information requested.20 With regard to foreign or international law enforcement officials, the Act is silent.21 Importantly, there is a distinction made between private OSPs and OSPs that serve the public—public-facing OSPs, the category under which Facebook, YouTube, and the like fall, are always subject to the provisions of the SCA and compliance is mandatory with regard to its rules that generally prohibit the disclosure of content.22

As was stated above, the SCA is silent on the matter of foreign or international law enforcement in favor of only explicitly allowing domestic law enforcement to request court actions to view stored data. In addition, longstanding US legal tradition holds that foreign law enforcement actions on its soil must be announced to and accepted by the government, even when a Mutual Legal Assistance Treaty (MLAT) exists between it and the foreign country conducting the investigative action. Because it would appear that the ICC would have to get help from US entities when searching for digital evidence removed and sequestered to an OSP’s server, we next have to analyze whether the ICC can effectively ask US legal and governmental institutions for aid when US-based OSPs remove potentially pertinent evidence posted online.

The most prevalent law governing US –ICC interactions since its adoption in 2002 is the American Servicemembers’ Protection Act (ASPA). In general, the ASPA sets out rules prohibiting US support for the ICC, including specific provisions against responding to a cooperation request from the ICC, transmitting a letter rogatory from the ICC, and using certain funds to aid the Court.23 Moreover, the Act bars the ICC and its “agents” from investigating within the US.24 Though those provisions seem to wholly remove the prospect of any potential aid from US institutions for the ICC, let alone help from the US in order to gain access to the servers of one of its companies, a last minute amendment by Senators Chris Dodd and Patrick Leahy (the “Dodd Amendment”) opens up the possibility of US aid in certain scenarios. The amendment states in relevant part:

[N]othing in this title shall prohibit the United States from rendering assistance to international efforts to bring to justice […] foreign nationals accused of genocide, war crimes or crimes against humanity.25

In addition, there is the potential for a Presidential waiver in order for the ICC to involve itself in an investigation in the US, though none of these waivers have been given out as of yet.26 This provision, on its face, allows US entities to cooperate with ICC investigations of non-US citizens and, as argued in Access Denied, opens up two interesting options for the ICC to work directly with US institutions on investigations, including when the Court needs access to an American OSP’s servers as a result of the removal of evidence posted online due to the OSP’s community guidelines. Those options are:

1. a court-to-court approach through a request for judicial assistance from the US District Courts, and

2. a court-to-country based, diplomatic approach.27

1. Requests for Judicial Assistance from US District Courts

Under Section 1782 of Title 28 of the U.S. Code, a section outlining proper assistance to foreign and international tribunals, foreign tribunals are permitted to send a letter rogatory to the district court attached to the location of the individual or entity they are pursuing, but only that court.28 The district court then is empowered to order the production of documents (amongst other rights), including for the purpose of foreign criminal proceedings conducted prior to formal accusation.29 Though the plain meaning of the Statute, which states that “foreign or international tribunals” can utilize this direct-to-court letter rogatory, should effectively qualify the ICC to utilize Section 1782, further proof of the right of the ICC to use this method was outlined in the Supreme Court case Intel Corp. v. Advanced Micro Devices, Inc., which established that the Statute may be used by any “interested person” and that the proceedings in question need not be imminent as long as they are “within reasonable contemplation.”30 This would seem to indicate that the Statute is interpreted by the US court system as broadly applicable and would allow for inquiries to be made during the preliminary examination phase.

The most significant benefits of addressing the issue of evidence being posted online and removed through a direct appeal to the US district courts are the timeliness and legitimacy factors, as well as removing the specter of noncooperation from the executive branch. American courts have among the best infrastructure of any legal system in the world, so the process will be as quick as possible and the affirmation of the relevance of the ICC investigation by an American court could lend it additional credence. Moreover, courts will look at an issue through a mostly legal lens (ideally), and, as such, political and security concerns will play less of a role.31 However, content data is not accessible through Section 1782 and the investigator would have to pinpoint the location of the data so that he or she could bring suit in that court, making it difficult to follow through on in practice.32 If content data is unavailable through this channel, it does not provide sufficient coverage to be a viable option through which to retrieve removed data.

2. Request for Judicial Assistance through Diplomatic Channels

Another potential route for the ICC to work directly with the American Government is through diplomatic letters rogatory sent to the State Department. Under 28 U.S.C. § 1781, the State Department is empowered to transmit letters rogatory to the addressed tribunal in the United States from foreign entities.33 Those are executed based on treaty obligations or “international comity and courtesy” and sent through the Justice Department’s Office of International Affairs (OIA). From there, OIA can proceed through the § 1782 process alluded to above, or it can use 18 U.S.C. § 3512 (the Foreign Evidence Request Efficiency Act or FEREA), which allows for the appointment of an Assistant US Attorney to act as an investigative intermediary and collect the necessary evidence.34 In addition, the Statute allows a federal judge to issue orders for appearance of witnesses or production of documents, and even allows federal judges to issue warrants to aid in international criminal investigations.35 FEREA differs significantly from § 1782 in that the request for assistance does not have to be filed in the specific court of jurisdiction and it opens up the possibility of access to content data as well as the metadata accessible under § 1782.36

The potential benefits for the ICC in utilizing letters rogatory through the State Department and the OIA are that it could extend the reach to content data, which would make it an actually viable option for the retrieval of data stored on servers located in the United States but removed from public view by OSPs for violating community guidelines. Rights under FEREA would need to be held to extend to international tribunals, as the Statute addresses requests from “foreign authorities” (though it is reasonable to believe that it would be, as criminal investigatory bodies are included under that moniker),37 but the extension of the right of foreign tribunals to access stored content data would essentially nullify the issue at hand. Despite that, these letters rogatory depend more heavily on the cooperation and discretion of the American executive branch, which does not bode well for their success rate, and international requests involving the discretion of multiple American branches of government tend to take significant time to complete. While diplomatic letters rogatory could potentially provide a direct solution to recovering removed digital data from OSP servers located in the United States, it would be necessary to wait for the right American administration to begin the process, the accused could not be American or from a US strategic ally, and the data in question would have to be stored in the US. In a situation which satisfies all of the above-mentioned conditions, they could be valuable.

B. Ireland

The public-facing OSPs that are most frequently at risk of removing pertinent digital evidence of crimes against humanity are all multinational organizations that host servers in multiple countries, despite being founded in the US. Therefore, some relevant data that is removed will be stored on servers in other countries necessitating involvement with nations outside of the US, and moreover the ICC could potentially look to partner with a state party to the Rome Statute that hosts a significant number of OSP servers to retrieve evidence, as those state parties would be more amenable to cooperation with the ICC than non-state parties. Ireland, a state party to the Rome Statute, is the EMEA headquarters for both Google and Facebook, as well as a host of other major OSPs. In 2006, Ireland passed the International Criminal Court Act (ICCA), which, in part, allowed for the issuance of warrants to access records or documents when the ICC requests state cooperation.38 It also has a law on the books, the Criminal Justice Act of 2011 (CJA), which imposes liability on officers or directors of OSPs that, out of willful neglect or a more culpable mens rea, delete or remove evidence from online access.39 On its face, Ireland seems like a reasonable venue for success in obtaining evidence stored on OSP servers.

However, the Irish legal setup leaves some gaps that make it ineffectual for use in attempting to retrieve data removed from public view. The CJA, primarily, does not list violations falling within the jurisdiction of the ICC among the crimes for which online evidence must be preserved,40 so it would be difficult to utilize coordinated threats under that Statute. Moreover, when OSPs have removed data pertaining to ongoing or potential ICC investigations, the process was either automated or driven by user complaints, not done at the direction or due to the negligence of employees at either company. Finally, while it is helpful that the ICCA calls for information sharing between Ireland and the ICC, and that channel should be explored where necessary, it only provides access to data stored on those servers located in Ireland and would not prove persuasive in an American domestic setting or any other national domestic setup. The root issue is the same as it was in the US context: access through national legal or governmental means alone is not sufficient to cover the totality of data stored on multinational OSP servers.

C. Mutual Legal Assistance Treaties

Although attempting to find a backdoor to solidify access to evidence posted online and removed from public access but stored on OSP servers through a national legal setup was unsuccessful at addressing the problem of data lost from public view in its entirety, perhaps the ICC could utilize existing bilateral MLATs where possible to access data where it was previously unable to. While the ICC cannot enter into MLATs or make original requests through an MLAT regime, as it operates on the traditional treaty basis of reciprocity, many states parties to the Rome Treaty do have existing MLATs with the United States. The ICC can ask any state party to request information on its behalf, and the Dodd Amendment would most likely legitimize any requests by the ICC pertaining to investigations of non-US citizens despite a provision of the ASPA that forbids the use of MLAT requests for ICC purposes.41 This method is particularly appealing where applicable because most major OSPs have policies that allow for freer information sharing through MLAT requests. Google, in particular, will produce the same information to a foreign agency requesting through the MLAT procedure as it would to a requesting US agency.42 Therefore, much like the diplomatic letter rogatory approach, the MLAT procedure would provide the ICC with the entirety of the data it seeks if it is successful.

Much like the purely national means of accessing data that has been removed by OSPs, the MLAT approach leaves gaps in the access available to the ICC. Significantly, these requests carry the most weight when they come from a state that referred a situation to the ICC themselves and is gathering their own evidence for a similar domestic criminal proceeding.43 This considerably narrows the scope through which MLAT requests could reasonably be expected to produce the desired evidence. In addition, many MLATs require permission from the state receiving the request before any information can be used for purposes beyond the scope of the treaty,44 so even though the Dodd Amendment opens up US MLATs for use by the ICC when investigating non-US citizens, it can likely still exercise discretion in whether or not to comply with the request. Much like the purely national means of access to data stored on OSP servers after being removed from public view, access through MLATs is effective only insomuch as a particular scenario allows it. When a state party has an MLAT with the appropriate country, self-referred, and is also looking for the same data as the ICC, and that data has been removed by an OSP and is being stored on the servers of a non-state party, the MLAT process provides a possible, albeit time intensive and discretionary, route to recovering the necessary data. Otherwise, it is likely less effective.

On the whole, the existing legal systems that govern national and cooperative bilateral legal systems do provide potential openings through which the ICC can access OSP servers in which removed pieces of evidence still reside. Given a situation where the content itself is removed from public consumption, the most effective means of accessing that stored data would be if that data happens to be stored in a country that is a state party to the Rome Statute, like Ireland. Short of that, when dealing with US firms and US servers, the most effective existing legal means to access stored data would be through diplomatic letters rogatory, which provide the most wide-ranging possible access despite the diplomatic process being lengthy and not completely reliable. If, as is the case with the Syrian conflict, Americans are involved and data is removed, there is a massive gap in the ability of the ICC to access data stored on US servers, as the Dodd Amendment only allows ICC action within the US when it is needed to investigate a foreign national. In addition, these potential solutions fail to address the fundamental issue of the problem presented: when data is removed, unless it is copied and reproduced somewhere else, the public at large is not able to identify it. Essentially, the existing legal processes through which the ICC can access data stored on OSP servers are only effective where the Court knew about the existence of the data beforehand and can pinpoint its location and the server it would have been backed up to. The question remains, therefore, as to what actions the ICC can take to prevent the loss of data as a result of blanket community guidelines established by OSPs in the first place.

IV. Alternatives to the State-Based Model: Multilateral, Bilateral, and Unilateral Options for the ICC to Protect and Maintain Evidence Posted Online

This section outlines and examines two potential solutions to the issue of losing digital evidence of war crimes to wide reaching community guidelines that are enforced by user submission or automated process. The first potential solution is the option for the ICC to seek to enter bilateral cooperation agreements with OSPs that host a significant amount of data from areas of investigation. More pertinently, however, the other solution examines projects to create a database of verifiable, protected digital evidence of human rights violations and suggests a potential ICC-run forum for online evidence collection.

A. Bilateral Cooperation Agreements with Online Service Providers

While the ASPA limits the possible ICC cooperation US governmental entities can undertake, there are no specific barriers to direct cooperative agreements between the ICC and private corporations.45 Superficially, a bilateral cooperation agreement between the ICC and the OSPs with community policies that risk the removal of pertinent digital evidence of war crimes, could potentially designate backup servers for such evidence. That would be an incredibly significant step in terms of evidence collection, and there are no legal barriers to such an agreement.

That being said, company policy is entirely a matter of discretion, and as such any agreement would be purely voluntary and unreliable.46 While such an agreement is possible, it has never occurred mostly because in would involve direct negotiation with companies that have widely variant approaches and policies, that all behave differently, and that each take a myriad of factors beyond the simple push for justice into account when making decisions. In addition, direct relationships between the Court and certain OSPs could give rise to questions of arbitrariness.47 This option also, much like direct letters rogatory to US district courts, limits access to metadata as opposed to content data because content data cannot be legally disclosed.48 Once again, metadata, while valuable in other contexts, is not helpful when the content is what needs to be recovered. Therefore, while, in theory, bilateral agreements between ICC and OSPs would be the most effective means through which the Court could guarantee access to data, they are unlikely to actually come to fruition.

B. Establishment of a Private Entity to Collect and Store Evidence Posted Online

The final solution proposed by this comment is one that is purely a preventative step, as opposed to working to preserve and/or access data once it is removed. Multiple organizations have undertaken relatively successful Internet or app-based projects which are focused on collecting and storing digital evidence of multiple crimes against humanity. The International Bar Association (IBA), for example, created the eyeWitness to Atrocities project, a mobile app purely for the purpose of collecting and securing online video evidence of crimes against humanity in response to difficulty obtaining information regarding Sri Lankan troop actions against Tamil prisoners.49 eyeWitness, while effective for its purpose, does not provide a great model for an ICC for a few reasons. First, it was only meant to record and store video footage. It is also its own separate entity incorporated in England and Wales, which would open up potential national legal issues with regard to privacy and corporate protections. eyeWitness also partnered with LexisNexis for use of its servers.50 If the ICC follows a similar path, it would certainly be helpful in terms of cost and logistics, but that could open up issues about the location of servers. Attachment to an OSP for use of its servers or the creation of a separate, private entity would necessitate some form of perpetual agreement between the ICC and that entity—whether that is possible, let alone likely, is unclear.

Perhaps a more robust model to follow stems from Benetech, a firm that created the Martus technology, which was a free, open source, public application that was built to collect, store, backup, and analyze data from activists and journalists documenting potential human rights violations. Essentially, it was a cloud-based service that allowed those documenting evidence of crimes against humanity to store and protect that evidence.51 The technology was active for upwards of fifteen years and was only deactivated due to fears that it was out of date, not for lack of use or need.52 The main obstacles to the ICC undertaking a project like this would clearly be cost and logistical—the Court would have to set aside significant amounts of its budget to undertake a project like this on its own.

The major drawback, aside from cost, to working to establish an ICC specific or a Court-connected private database of digital evidence is that, generally, social media platforms have more name brand value.53 It would take significant time and advertising to grow the general knowledge of a new ICC program. If that is the main issue, however, and the upside is a database through which the ICC can avoid dealing with online service providers entirely, then the potential benefit far outstrips the cost needed to get there.

V. Conclusion and Policy Recommendation

In sum, the best possible solution for the ICC to combat the problem of OSP community guidelines indiscriminately removing from public view caches of potentially pertinent digital evidence would be to work to establish its own private forum for the uploading, safekeeping, and analysis of evidence of war crimes posted online. The best model for this would be the Martus model, but, if the Court does not have the time, funding, or logistical knowledge to establish its own application, an expansion of the IBA model with a perpetual cooperation agreement would be the next best option. While those are explored further, maintaining dialog with OSPs regarding the effect of their policies on the Court could potentially move certain firms, especially those like Facebook which have received copious negative publicity for its handling of such evidence on its site. But changing the policies of OSPs is not a guaranteed solution. Short of all of that, the best method for reaching evidence removed from public view but stored on OSP servers would be to go through states parties to the Rome Statute, or through diplomatic letters rogatory in the case of servers located in the United States.

Endnotes — (click the footnote reference number, or ↩ symbol, to return to location in text).

1. 1.

   Peggy O’Donnell, Alexa Koenig, Camille Crittenden & Eric Stover, UC Berkeley HRC, Beyond Reasonable Doubt: Using Scientific Evidence to Advance Prosecutions at the International Criminal Court (Jan. 7, 2013), available online. ↩

2. 2.

   Róisín Á Costello, International Criminal Law and the Role of Non-State Actors in Preserving Open Source Evidence, 7 Cambridge Int’l L.J. 268 (Dec. 1, 2018), paywall, doi. ↩

3. 3.

   Id. ↩

4. 4.

   Id. ↩

5. 5.

   Avi Asher-Schapiro, YouTube and Facebook are Removing Evidence of Atrocities, Jeopardizing Cases Against War Criminals, The Intercept, Nov. 2, 2017, available online. ↩

6. 6.

   ICC Approves Probe into Myanmar’s Alleged Crimes Against Rohingya, Al Jazeera, Nov. 14, 2019, available online. ↩

7. 7.

   Community Standards: Violence and Criminal Behavior, Facebook, available online (last visited May 29, 2020). ↩

8. 8.

   Community Standards: Safety, Facebook, available online (last visited May 29, 2020). ↩

9. 9.

   Id. ↩

10. 10.

    Asher-Schapiro, supra note 5. ↩

11. 11.

    Id. ↩

12. 12.

    Id. ↩

13. 13.

    Id. ↩

14. 14.

    Id. ↩

15. 15.

    Julia Carrie Wong, Michael Safi & Shaikh Azizur Rahman, Facebook Bans Rohingya Group’s Posts as Minority Faces “Ethnic Cleansing,” The Guardian, Sep. 20, 2017, available online. ↩

16. 16.

    Violent or Graphic Content Policies, YouTube, available online (last visited May 29, 2020). ↩

17. 17.

    Hate Speech Policy, YouTube, available online (last visited May 29, 2020). ↩

18. 18.

    Asher-Schapiro, supra note 5. ↩

19. 19.

    Tommy Umberg & Cherrie Warden, The 2013 Salzburg Workshop on Cyber Investigations: Digital Evidence and Investigatory Protocols, 11 DEESLR 128 (Nov. 21, 2014), available online, doi. ↩

20. 20.

    Id. ↩

21. 21.

    Id. ↩

22. 22.

    Id. ↩

23. 23.

    Alexa Koenig, Keith Hiatt & Khaled Alrabe, Access Denied? The International Criminal Court, Transnational Discovery, and The American Servicemembers Protection Act, 36 Berkeley J. Int’l L. 1 (May 2018), available online. ↩

24. 24.

    Id. ↩

25. 25.

    Id. ↩

26. 26.

    Id. ↩

27. 27.

    Id. ↩

28. 28.

    Id. ↩

29. 29.

    Id. ↩

30. 30.

    Id. ↩

31. 31.

    Id. ↩

32. 32.

    Id. ↩

33. 33.

    Id. ↩

34. 34.

    Id. ↩

35. 35.

    Id. ↩

36. 36.

    Id. ↩

37. 37.

    Id. ↩

38. 38.

    Costello, supra note 2. ↩

39. 39.

    Id. ↩

40. 40.

    Koenig, Hiatt & Alrabe, supra note 23. ↩

41. 41.

    Id. ↩

42. 42.

    Umberg & Warden, supra note 19. ↩

43. 43.

    Koenig, Hiatt & Alrabe, supra note 23. ↩

44. 44.

    Id. ↩

45. 45.

    Id. ↩

46. 46.

    Costello, supra note 2. ↩

47. 47.

    Koenig, Hiatt & Alrabe, supra note 23. ↩

48. 48.

    Id. ↩

49. 49.

    Costello, supra note 2. ↩

50. 50.

    Id. ↩

51. 51.

    Overview, Martus, available online (last visited May 29, 2020). ↩

52. 52.

    Collin Sullivan, Martus Sunset, Benetech (May 15, 2018), available online. ↩

53. 53.

    Costello, supra note 2. ↩