An offer they can’t refuse: Israel’s ‘informal’ censorship meets Facebook’s compliance

Who will keep tabs on the censorship of Israel’s internet? The state attorney’s Cyber Unit “asks” internet companies to remove allegedly problematic content – requests that are overwhelmingly granted. But there’s no documentation and no supervision. No one knows what’s being censored, why and who calls the shots.

Who will keep tabs on the censorship of Israel’s internet? The state attorney’s Cyber Unit “asks” internet companies to remove allegedly problematic content – requests that are overwhelmingly granted. But there’s no documentation and no supervision. No one knows what’s being censored, why and who calls the shots.

Who will keep tabs on the censorship of Israel’s internet? The state attorney’s Cyber Unit “asks” internet companies to remove allegedly problematic content – requests that are overwhelmingly granted. But there’s no documentation and no supervision. No one knows what’s being censored, why and who calls the shots.

Processing: Noam Tamari | Photos: Reuters

Viki Auslender

in collaboration with

November 26, 2020

Summary

F

or many people, including, no doubt, public figures and celebrities of various kinds, the power in the hands of the 26 employees of the Cyber Unit at the Office of the State Attorney is tantamount to a dream that can only be described as too good to be true: The power to censor the internet.

In this, the 21st century, that means made-in-China superpowers to delete posts, websites, blogs, social media pages, tweets, and the like. Digital memory, as we all know, doesn’t forget a thing, with long-forgotten moments of utter despair or offensive publications simply pushed to the back pages of the search engines at best.

Last year, the Cyber Unit submitted no fewer than 19,606 requests for content removal to internet providers and internet companies operating in Israel, mainly Facebook, Twitter and Google. A staggering 90 percent of them were voluntarily granted. The full extent of this willingness will be examined in further detail later on in this piece.

Only 56 of the cases last year resulted in criminal proceedings of any kind, with the unit having to legally force the removal of the content. In 2018, the number of requests totaled 14,283 (86 percent were voluntarily granted by the internet companies); in 2017, it was 12,351 (85 percent voluntarily granted); and in 2016, there were just 2,241 requests (70 percent voluntarily granted). Recent years, in other words, have seen a significant increase in both the unit’s appetite and the internet companies’ willingness and compliance.

On the face of it, there’s nothing to gripe about. The Cyber Unit was established in September 2015 with the stated purposes of “dealing with crime and terrorism in cyberspace.” Any mention of the word “terrorism” immediately brings us into the realm of state security, and everyone zips their lips and the doors are closed, even before another word is said. The picture, however, is more complex, and there are many reasons for this. There are two main ones – the first, a lack of transparency; and the second, a controversial authority that would rather meander through gray areas than follow the legal path.

The voluntary route that the unit’s activities have followed is backed up by a procedure that has become commonplace in recent years: the unit receives requests to remove content from military and civilian security organizations and has the option of going down a legal path, seeking court orders and getting the criminal justice system involved. Alternatively, it can simply use its own channels directly to contact Facebook – which, in the vast majority of the cases, is the platform in question – and to “advise” the company to remove the content willingly. The unit almost always goes for the second option.

To what extent are we dealing with goodwill on the part of Facebook regarding the substance of the matter at hand, or goodwill designed to avoid further entanglement in government legislation dubbed the “Facebook Law,” which was vigorously promoted until two years ago by then ministers Ayelet Shaked and Gilad Erdan, and which hung like a sword over the social network’s head? There are movies in which “advice” of this kind is known as an offer you can’t refuse.

And this is where we also find an open window in terms of the lack of transparency and the potentially slippery slope for freedom of expression that could very well follow: What content does the unit choose to remove from the web, courtesy of Facebook and other internet companies? What kind of content – security-related or perhaps political and personal – is censored?

We would have been happy to provide a handful of concrete and significant examples to demonstrate the praiseworthy activity of the unit, but the only answer we got was that there is no answer. The Cyber Unit doesn’t document a thing. Nothing at all. Because keeping records, the unit tells Shomrim, “gives rise to a host of legal and technological difficulties.”

It follows, therefore, believe it or not, that in the absence of documentation, there is no way of learning about or monitoring the unit’s activities when it comes to content that is voluntarily removed by the service providers and internet companies – where and why the unit draws red lines, what do its chiefs deem permissible to say and what’s forbidden; and there’s no way to judge or examine whether its decisions were justified, or at least proportionate.

What content does the unit choose to remove from the web, courtesy of Facebook and other internet companies? What kind of content – security-related or perhaps political and personal – is censored? The Cyber Unit doesn’t document a thing. Because keeping records, the unit tells Shomrim, “gives rise to a host of legal and technological difficulties.”

The Trojan Horse display at Tel Aviv University during National Cyber Week in 2016. Photo: Reuters

In response, for example, to a request for information submitted by the Movement for Freedom of Information and the Clinic on Human Rights in Cyberspace at Jerusalem’s Hebrew University, it was reported that over the past three years, the Israel Police has submitted 80 content-removal requests to the Cyber Unit, with 35 of them passed on to the internet companies, and 28 of those 35 requests granted. What was the content that was asked to be removed? The unit doesn’t publish, document, or share this information, even though this isn’t a matter that comes under the vague and broad definition of “state security.” The unit is willing to disclose nothing more than a general breakdown, according to which 76 percent of the requests for voluntary removal they passed on in 2019 were related to offenses involving terror organizations, and 22 percent were incitement offenses. Examples? There are none.

Doesn’t the public have the right to know? Or shouldn’t external entities at least have the option of reviewing the unit’s decision-making process when it comes to curbing freedom of expression online? The Office of the State Attorney thinks not. Others will beg to differ.

“State intervention in public discourse on the internet is intervention in the freedom of expression of the citizens of the state and their right to consume information,” says Dana Yaffe, a lawyer who heads the Clinic on Human Rights in Cyberspace. “It’s a significant violation of basic rights. Even if there are instances in which such intervention is justified, actions of this kind must be open to public oversight and review,” she adds.

“They keep saying there aren’t any records and I don’t believe them, because clearly they can be traced,” notes Advocate Rachel Edri-Hulta, the CEO of the Movement for Freedom of Information. “After all, these things don’t just happen out of the blue. They follow in the wake of requests, of emails. They also say that allowing us access to the requests could undermine state security, and we don’t like that argument. These things have a time and place in which they could be offensive and dangerous – and there’s a time and place in which they aren’t so. It depends on the context in which they’re found.”

The court will decide

The Cyber Unit was established five years ago, in the shadow of the so-called Knife Intifada, a wave of terror attacks that started in the fall of 2015 and continued into the first half of 2016, with the perception being that the increased violence was the result, largely, of ever-increasing incitement on social media. The unit’s activities, therefore, are focused on three main functions – to provide professionally guidance to attorneys and enforcement agencies on issues such as digital evidence and wiretapping, to manage criminal cases in the field of computer and information offenses, and to act in the fields of alternative enforcement.

Since the adoption of the new Counter-Terrorism Law in 2016, and the Law on Authorities for the Prevention of Committing Crimes Through Use of an Internet Site in 2017, the Cyber Unit has been equipped with the tools to act along the classic lines of applying for a court order to remove content that promotes prostitution, pedophilia, drug trafficking, illegal gambling or terror organizations.

But what happens when offenses occur outside the borders of the state for example, where the unit has no jurisdiction? Or when the unit believes it doesn’t have the time to act through the legal channels? Or when the alleged perpetrator is anonymous? This is where the unit implements its unique censorship mechanism, its so-called “alternative enforcement,” a concept formulated by Dr. Haim Vismonsky, one of the founders of the unit and its director since 2015.

Dana Yaffe: “When there’s no documentation, there’s no way of knowing how the state handles these requests, what, in its view, constitutes content that requires removal, whether it acts only when the matter at hand is a criminal offense, whether there are biases in play and selective enforcement in favor of a particular group or idea."

Attorney Dana Yaffe, The Clinic on Human Rights in Cyberspace

This type of enforcement, as the unit views things, is designed to deal with the dynamic reality of cyberspace, using a tool that isn’t bound by the limitations of the slow and laborious formal legal path. Under this alternative method, the Cyber Unit in the service of the State of Israel doesn’t carry out the enforcement itself but consciously passes on the responsibility to the internet content companies, in a non-binding and non-coercive manner of sorts – not in the legal sense at least. According to this approach, the state doesn’t need the power of the law, because it isn’t really acting with specific legal authority; it’s merely advising, “nagging a little,” and bringing the matter to the attention of the companies. The state itself ostensibly does nothing.

One example that sources at the Cyber Unit were kind enough to recall, nevertheless, took place in July 2017, following a terror attack in the Halamish settlement, in which three Israelis were killed. At the time, the unit contacted Facebook and asked the company to remove a viral post that contained the attacker’s “last will and testament.” The post, the unit claimed, had become a focal point for glorifying the terror attack and encouraging others.

At first glance, this would appear to be an example of exactly the kind of security-related incident for which the unit was established (even if there’s no entity to review the decision-making on these issues either). But what about civilian issues? In February 2019, for example, a television report revealed that the Israel Police was appointing “shaming officers” to help any police officer or volunteer who felt harmed by online publications concerning his or her work. The procedure, which went into effect two years earlier, created a reporting mechanism for police complaints regarding offenses such as insulting a public official, invasion of privacy, or intimidation. Complaints deemed worthy by the shaming officer were then forwarded to the Cyber Unit, which decided whether or not to pass them on to the various internet companies and have them remove the content – voluntarily, of course. The data in this regard is anyone’s guess.

Rabea Eghbariah, an attorney from Adalah - The Legal Center for Arab Minority Rights in Israel, is vocal in his criticism of the voluntary censorship route developed by the Office of the State Attorney. “This is a total collapse of the separation of authorities, because the Office of the State Attorney alleges and decides, ostensibly, to file an indictment, and also judges and asks for the removal itself,” Eghbariah says. “What we’re dealing with here is the outsourcing of the governmental authority and the privatization of the judicial authority. The entire mechanism becomes an internal mechanism for the Office of the State Attorney that’s presented as a way of cooperating with the companies.”

In 2019, Adalah, together with the Association for Civil Rights in Israel, filed a petition with the High Court of Justice against the alternative enforcement mechanism employed by the Cyber Unit. Their arguments encompass all the unit’s activities, charging that the unit is operating without any legal authority, that it’s making decisions concerning freedom of expression offenses without any court proceedings, that its activities are a deathblow to freedom of expression, and that it’s violating the principle of the right to fair legal proceedings.

Rabea Eghbariah: "This is a total collapse of the separation of authorities. The Office of the State Attorney alleges and decides, ostensibly, to file an indictment, and also judges and asks for the removal itself. What we’re dealing with here is the outsourcing of the governmental authority and the privatization of the judicial authority.”

Attorney Rabea Eghbariah. Photo: Mati Milstein, Adalah Center

In rejecting the petition, the state argued that it doesn’t act without authority because it doesn’t exercise any authority at all. According to the state, the unit’s requests to remove content are not “demands,” and that the decision whether or not to remove content is entirely at the discretion of the online platforms.

“One should always remember that there is a formal route,” says Eghbariah. “The problem is that the Law on Authorities for the Prevention of Committing Crimes Through Use of an Internet Site affords limited powers and doesn’t allow for the removal of specific content. So instead of the formal rule of law, they’ve gone for an alternative rule of law. They’ve said: We’ll come up with an alternative form and alternative legal principles and make an alternative direct approach to the internet companies and ask them to remove content voluntarily, thus circumventing the court.”

The High Court heard the petition last August, and instead of rejecting it, as the state demanded, it instructed the state to complete its arguments regarding the source of its authority to take the voluntary route with the internet companies, and to add a review of similar legislation or arrangements in various countries around the world.

“Like a black box”

The Movement for Freedom of Information, which requested to join the petition filed by Adalah and the Association for Civil Rights, is less troubled by the issue of the source of authority and more concerned about the matter of transparency. “Do the work, first and foremost, transparently,” says the movement’s director, Rachel Edri-Hulta. “We’d like to get to the bottom of the relationship between Facebook and the State of Israel – if indeed there is one. We want to see how many [of the requests for the removal of content] concern issues that touch on state security, and how many, for example, are related to the defaming of judges.”

(Note, for the sake of full transparency, that Shomrim is working with the Movement for Freedom of Information on several additional petitions, such as a petition to receive the transcripts from the discussions of the government’s Corona cabinet, and a petition against the Civil Service Commission to receive information on vacant positions in the service or positions that are filled currently with temporary staffers.)

Rachel Edri-Hulta: “They keep saying there aren’t any records and I don’t believe them, because clearly they can be traced. After all, these things don’t just happen out of the blue. They follow in the wake of requests, of emails.

Attorney Rachel Edri-Hultam, The Movement for Freedom of Information. Photo: Ilia Malenikov

Edri-Hulta has good reason to insist on the issue of transparency. Since the establishment of the Cyber Unit, numerous efforts and inquiries have been made to obtain information on the content the state removes, and to require it to reveal the nature of its activities, beyond the general numbers that it releases at the end of each year via the Office of the State Attorney – numbers that, in any case, don’t reveal anything other than the scope of its activities. But the unit has consistently refused to cough up.

It claimed, initially, that it didn’t have the resources to accumulate the data or the technical means to record it. Thus, in response, for example, to a petition filed last December by publicly-owned media organization Shakuf, which requested details on the content that had been removed during the period 2017-2018, the state noted, inter alia, that the request placed “an unreasonable burden on the authority.” The state went on to argue subsequently that it doesn’t record the information because the publications themselves are a violation of the law and it has no interest in saving them, adding that in any event, sharing such information would undermine state security, which trumps freedom of information.

As far as the state is concerned, the general numbers it publishes, and which appear here in this report, constitute sufficient transparency. In the same vein, the state asked the High Court to reject the Movement for Freedom of Information’s request to join the petition filed by Adalah and the Association for Civil Rights.

“Removing content without transparency and without authority raises concerns of disproportionate prejudice,” adds Eghbariah. “In our experience, the first victims are the Arab citizens in the country and the Palestinians. That’s why we’ve been skeptical about this practice, particularly because the Cyber Unit reports that the vast majority of the content it removes comes under the umbrella of terrorism and incitement. But who determines what terrorism is and what incitement is? They could include a very wide range of legitimate criticism of the state. There are feed lines going into the unit, and it’s like a black box."

Dana Yaffe, from the Clinic on Human Rights in Cyberspace, concurs: “When there’s no documentation, there’s no way of knowing anything at all,” she says. “There’s no way of knowing how the state handles these requests, what, in its view, constitutes content that requires removal, whether it acts only when the matter at hand is a criminal offense, whether there are biases in play and selective enforcement in favor of a particular group or idea, or whether language and culture gaps lead to errors towards a particular population sector”

In their view, as long as there’s no transparency, it’s impossible, too, to blindly accept the state’s claims that it isn’t forcing the removal of the content, and that the tens of thousands of requests it submits nonstop to the internet companies are nothing more than good-natured and voluntary cooperation and not the exercising of any authority. “It’s clearly being conducted in a specific political sphere and should be open to review and criticism at the very least,” says Adalah’s Eghbariah. “The attempt to portray the issue as ‘advice’ to internet content companies, with no balance of power at play, is inaccurate; advice requires a degree of authority too.”

The Cyber Unit offered Shomrim the following response: “The Cyber Unit at the Office of the State Attorney doesn’t actively seek out illegal content on the internet, but merely reviews the approaches it receives from other administrative authorities. The administrative authorities are the ones that locate the publication on the internet in the first place. They document it, in keeping with their internal procedures, formulate an orderly and reasoned approach, and pass on the information to the Cyber Unit. It should be noted that documenting content, as aforementioned, raises a variety of legal and technological difficulties.

“Regarding the voluntary track, the unit operates in keeping with an approved procedure. Regarding the formal track, the unit operates in keeping with the powers laid out in the law. The unit’s activities on the voluntary tracks are currently the subject of a High Court petition.

“The unit also deals with publications that express threats or incitement or a gross violation of privacy against public officials. The unit’s activity in this area constitutes less than 1 percent of the total number of offensive publications it handles. Most of the publications with which the unit deals in this area are security-related publications.”

Background | Supervising the internet: Wanted – responsible adult, but clear definitions too

In May 2017, internal Facebook documents outlining the company’s secret content-censorship policies were leaked and published in the British newspaper, The Guardian. The world was up in arms, enraged by the very idea that Mark Zuckerberg and his friends could decide for themselves where to draw the line between a post in bad taste and poisonous or inciting hate speech, and thereby censor the social network.

In Israel, a Knesset committee convened and the sparks flew, with parliamentarians demanding that Facebook report for questioning to explain and outline the reasoning behind its policies, and discussing among themselves how to monitor the activities of the internet giant. What they didn’t know was that the state and Facebook were already involved in a very close reciprocal relationship in terms of monitoring the social network.

The first glimpse into this relationship was provided by former justice minister Ayelet Shaked a year earlier. In June 2016, at a conference in Hungary on the fight against online incitement and shaming, Shaked said that the social networks responded positively to requests from the Israeli Justice Ministry to remove content. “Israel works with Google, Facebook and Twitter to remove incitement,” the then-minister said. “We have a 70 percent success rate.”

The relationship between the Israel and Facebook, as well as with other online content platforms such as Google and Twitter, is the result of a rude awakening from the euphoria of the internet and information technology revolution. Yes, the internet has indeed provided billions of people around the world with the opportunity to express themselves without any barriers or costs, but the fact that the structure of the platforms themselves encourages the radicalization of the discourse is also plain to see. Thus, the ugliest sides of freedom of expression – hate speech, incitement, violence, racism, scare tactics, and the dissemination of fake information – have been given a stage on which to perform and shout out loud for all to hear.

For years, Western nations refrained from policing the internet themselves, primarily due to concerns about that slippery slope when it comes to regulating freedom of expression. Instead, they demanded that the internet companies, huge commercial enterprises, do the policing – without offering guidelines or clear definitions.

Eventually, the pressure paid off, and to preserve their power, the internet companies fell in line with these demands and started to adopt an inconsistent and subjective regime to curb the discourse. To cope with the huge volumes, they “trained” algorithms to locate and remove content, and also began employing thousands of content sifters – contract workers who undergo brief training from company representatives and sometimes don’t even speak the original language of the content they oversee – to rule on subjective questions related to freedom of expression.

This, of course, didn’t go down very well. The idea that the founders of Facebook or Twitter would impose their moral relativism and censor the internet enraged users, governments, journalists, academics, right-wingers and leftists, defenders of freedom of expression, fringe groups, fans of conspiracy theories, and alt-right groups. No one was happy – not with the limitations imposed on freedom of expression and not with the way in which the lines were drawn.

Left with no choice, governments were forced to try to come up with an enforcement system for policing the internet. In Germany, France, Sweden and Norway, for example, laws have been passed to require internet companies to remove content deemed prohibited within a stipulated period of time, or pay a hefty fine. In Britain, there’s even a mechanism under which internet companies can be deemed criminally liable for failing to act promptly.

In Israel, a legislative initiative known as the Facebook Law, designed to give the government extensive powers to censor content on social networks, and promoted by former ministers Shaked and Erdan, was halted at the last minute in the summer of 2018. And into this regulatory space stepped the Cyber Unit, which doesn’t operate as part of the state’s attempt to determine the boundaries of free speech on the internet via legislation, but via skillfully leveraging the inconsistent, fuzzy and gray censorship regime of the internet companies instead. By means of this unit, Israel is dabbling in the world of state censorship, not to mention the subjective interpretation it often ascribes to terms such as incitement, hate speech, insulting a public official, or terrorist activity.

This is a summary of shomrim's story published in Hebrew.
To read the full story click here.