Science

When governments censor websites and block messaging apps like Telegram, here's where to turn for proof

As network filtering and censorship technology becomes easier to obtain and use, data collected by an organization called Open Observatory of Network Interference is helping hold governments to account.

Data collected by watchdog Open Observatory of Network Interference is helping hold governments to account

An Iranian man demonstrates that he can't use the messaging app Telegram in Tehran early January, which was temporarily blocked after anti-government protests erupted in Iran at the end of last year. (Abedin Taherkenareh/EPA-EFE)

In Iran, use of the messaging app Telegram has officially been banned. 

For some 40 million Iranians, Telegram has been an integral part of daily life, a place to talk with friends and family beyond the reach of government censors. Which is why, after anti-government protests broke out in the final days of 2017, the government instructed the country's internet service providers to implement temporary controls that would make Telegram harder to use — before outright banning its use this month.

Anecdotal reports are one thing. But to understand how, exactly, Telegram was being blocked — and to what extent in different parts of the country — researcher Mahsa Alimardani turned to technical data gathered by a watchdog group called the Open Observatory of Network Interference, or OONI.

It's about holding government officials to the fire and keeping their commitments to protect free expression clear.- Peter Micek, general counsel, Access Now

The organization collects evidence of internet censorship in more than 200 countries, with help from as many as 50,000 volunteers — often activists and human rights defenders but also regular citizens — who run OONI's internet measurement testing apps each month. The tests look for telltale signs of what people in the digital rights community call "information controls" — the myriad techniques used to block websites and apps or make them so slow to load they become impossible to use.

During the Iran protests, having that data "was really valuable to show what the government was doing," said Alimardani, who leads some Iranian digital rights programs for the advocacy organization ARTICLE 19 and is a doctoral student at the Oxford Internet Institute. 

"It's just good to have this kind of empirical evidence to supplement what's coming from anecdotal evidence," she said.

Angry over economic problems, Iranian students demonstrate at the University of Tehran December 30, 2017. After temporarily blocking access to the messaging app Telegram in January, its use was officially prohibited last month. (STR/AFP/Getty Images)

As network filtering and censorship technology become easier to obtain and use — and incidents of blocking, filtering, or full internet shutdowns increase — data collected by OONI is playing an increasingly important role in holding governments to account.

The goal, says Arturo Filasto, the project's lead and founder, is to complement the on-the-ground work of local partners with some "solid technical, undeniable evidence that these sorts of controls are in place."

'Holding government officials to the fire'

OONI was formed in 2011 as part of the Tor Project, a non-profit group that developed the Tor browser, which anonymizes users' web-browsing activity. That makes it ideal for circumventing censorship but also a prime target for government censors. 

After Tor users reached out for help, Filasto hacked together some simple software for the users to run. The data collected helped the Tor team understand precisely how Tor was being blocked, and that soon became a project of its own. Today, six people work on the core OONI team full-time.

Mahsa Alimardani, who leads some Iranian digital rights programs for the advocacy organization ARTICLE 19 and is a doctoral student at the Oxford Internet Institute, has used OONI data as part of her work. (Mahsa Alimardani)

All of the data collected by OONI's measurement software — called probes — is stored in a publicly accessible database, where anyone can go to understand what's being blocked, filtered, or throttled in a particular country, and how. That data can be used to track the evolution of information controls over time or link censorship with political events like elections and protests. 

For example, OONI tracked the censorship of news and human rights groups during 2016 political protests in Ethiopia and the blocking of social media both before and after Uganda's presidential election that same year. In Malaysia, news outlets that reported on a 2016 scandal involving the prime minister were also blocked. And, much like Iran, Russia has moved to block the use of Telegram in recent weeks.

Ethiopian protesters run from tear gas fired by police during Irreecha, the thanksgiving festival of the Oromo people, in Bishoftu town October 2, 2016. In response, the government restricted access to social media, news and human rights websites throughout the year. (Tiksa Negeri/Reuters)

The data has also proven enormously valuable to researchers working to document censorship around the world. The University of Toronto's Citizen Lab released a report at the end of April — relying, in part, on OONI data — about  the use of internet filtering technology from a Canadian company called Netsweeper. The researchers say Netsweeper technology is used in 10 countries to censor everything from political criticism to LGTBQ resources. 

"It's about holding government officials to the fire and keeping their commitments to protect free expression clear," says Peter Micek, the New York-based lawyer for the global internet access advocacy organization Access Now, which also relies on OONI data. 

AccessNow has been tracking the rise of internet shutdowns through its KeepItOn Coalition, which reported 104 internet shutdowns in 20 countries last year.

Censors getting smarter

Mapping instances of internet censorship is far from an exact science, and OONI's work is not without challenges.

Maria Xynou, who has handled community outreach and research for OONI since 2016, says the team has spent years trying to make its measurement software as easy as possible for less tech-savvy users — namely, the human rights advocates, lawyers and journalists who are most interested in running tests. 

OONI staff and community members at a Berlin hackathon in November, 2016. (OONI)

A mobile app for iPhone and Android released last year significantly expanded their coverage, and has helped OONI make more comprehensive observations. The nature of the internet is such that traffic isn't always routed along the same path. Sometimes a connection might route along a path where access is blocked, and other times not. Some internet service providers might block different sites and services than others, even in the same city. What's accessible can change depending on the IP address a person is connecting from.

It can also be hard to distinguish between instances of confirmed, intentional censorship, and weird network anomalies, even with technical evidence in hand. Some censors have noticed this too, and have eschewed the traditional block pages that are often displayed in place of the requested website in favour of more subtle tampering — making pages appear to load slowly, or not at all — so that it is harder to tell that censorship is actually taking place

"If more censors were to do something of that sort, I think our job would be a bit harder," Filasto says. "Because at that point it's pretty hard to distinguish what is intentional government mandated censorship from just some form of networking issue." 

People release paper planes, symbol of the messaging app Telegram, during an April 30, 2018, rally in Moscow in protest of a court decision to block the app. Telegram has resisted demands from the Russian government to make it easier for local law enforcement to read users' messages. (Tatyana Makeyeva/Reuters)

At the same time, OONI is trying to make things even easier for users. Later this year, Filasto expects the team to begin work on an automated censorship alert system. When OONI's systems detects the possibility of new information controls in a country of interest, people can be notified using something like a tweet, email, or push notification — which should make it even easier for the average person to track new cases of censorship in places that matter to them.

"I feel like it's a very empowering thing," says Xynou. "You don't need to trust your ISP to protect your rights. You don't need to trust your government. Instead, you as a citizen can inspect your own network and see what's happening."

ABOUT THE AUTHOR

Matthew Braga

Senior Technology Reporter

Matthew Braga is the senior technology reporter for CBC News, where he covers stories about how data is collected, used, and shared. You can contact him via email at matthew.braga@cbc.ca. For particularly sensitive messages or documents, consider using Secure Drop, an anonymous, confidential system for sharing encrypted information with CBC News.