No Access: LGBTIQ Website Censorship in Six Countries

Today, in collaboration with OutRight Action International and the Citizen Lab, we are excited to share our new research report, “No Access: LGBTIQ Website Censorship in Six Countries”, which examines the blocking of LGBTIQ websites in Indonesia, Malaysia, Iran, Russia, Saudi Arabia, and the United Arab Emirates (UAE).

READ FULL REPORT

Annotated Bibliography

Below we share some of our key research findings.

Summary of findings

We joined forces with OutRight Action International and the Citizen Lab to examine the blocking of LGBTIQ websites in six countries: Indonesia, Malaysia, Iran, Russia, Saudi Arabia, and the United Arab Emirates (UAE). We selected these countries because they are (a) known to serve block pages (i.e., pages that website visitors may see when access is restricted), which enable us to automatically confirm the blocking of LGBTIQ websites, and (b) known to censor LGBTIQ related content, based on prior research.

We adopted a mixed methods research approach, combining OONI network measurement analysis with interviews and literature research. Specifically, we analyzed OONI measurements collected from these six countries to detect the blocking of LGBTIQ websites, and to examine how their blocking varies across AS networks in each country (if at all). The timeframe that we selected for OONI data analysis was June 1, 2016 to July 31, 2020. To examine the impact of online LGBTIQ censorship, OutRight Action International and the Citizen Lab interviewed LGBTIQ communities in the six countries.

Our technical findings, along with information on the criminalization of LGBTIQ related activities, are summarized for each country in the following tables.

Image: Table 1: Summary of findings

Image: Table 2: Summary of findings

Overall, through OONI data analysis we found:

• Iran blocks the highest number of LGBTIQ URLs in our test lists. Out of the six countries, the highest instance of LGBTIQ URL blocking was seen in Iran, where 75 unique LGBTIQ URLs were detected as blocked. In Iran, we also observed the blocking of www.outrightinternational.org, the website of OutRight Action International, one of this report’s authors. In second place is the UAE where 51 unique LGBTIQ URLs in our test lists were found blocked.

• Russia has the highest number of networks that block LGBTIQ websites. Specifically, we detected the blocking of LGBTIQ websites on 172 distinct AS networks. Iran has the second highest prevalence of blocking, with LGBTIQ websites being blocked on 84 AS networks. In Indonesia, LGBTIQ websites were blocked on 43 AS networks, while in the UAE, LGBTIQ websites were found blocked on only 3 AS networks. We should note, however, that these figures also reflect the diversity of each country’s ISP market.

• Saudi Arabia has the highest percentage of LGBTIQ website blocking consistency. The highest blocking consistency was found in Saudi Arabia, where most LGBTIQ URLs that we tested were found blocked more than 75 percent of the time. Blocking consistency refers to the amount of times that we found each LGBTIQ URL to be blocked in each country in comparison to the total amount of times tested throughout our analysis period. The more times a URL is found blocked in comparison to the times that it was tested, the higher its blocking consistency. Our research across the six countries found variation in blocking consistency.

• LGBTIQ websites on “culture and community” were blocked most often. In Indonesia, Iran, Malaysia, and Saudi Arabia, the most frequently blocked LGBTIQ websites were those that belong under the “Culture and Community” category. These are websites that aim primarily to create a sense of community among LGBTIQ individuals, as well as provide information about art and culture. This is not the case in Russia, however, where LGBTIQ websites under the “News Media" category instead presented the most blocking, while in the UAE, most of the LGBTIQ websites found to be blocked were no longer operational (categorized as “404 Not Found”).

• Variation in the blocking of internationally-relevant LGBTIQ websites vs. locally-relevant ones. All six countries blocked LGBTIQ websites that are internationally-relevant and meant for an international audience, including popular dating websites (e.g., www.grindr.com), news media sites (e.g., www.advocate.com), and websites of organizations defending LGBTIQ human rights around the world (e.g., ilga.org). In Malaysia and Indonesia, all local LGBTIQ websites tested (e.g., queerlapis.com and suarakita.org), however, were accessible during our analysis period, and therefore, it appears that both countries block internationally-relevant LGBTIQ websites only. In contrast, Iran, Russia, Saudi Arabia, and the UAE blocked access to several local and regional LGBTIQ sites, in addition to blocking internationally-relevant LGBTIQ sites.

• Variation in the implementation of internet censorship across networks in the six countries. Iran, for example, appears to have a uniform censorship apparatus, as most ISPs not only block the same websites, but they also use the same set of censorship techniques. Nevertheless, Iranian ISPs appear to alternate between blocking and unblocking LGBTIQ URLs over time. In Indonesia, on the other hand, despite the Communications and Information Technology Ministry’s (Kominfo) Trust Positif censorship program, we observed differences in LGBTIQ URLs blocked on different networks. This variation signals that different ISPs block access to different websites at different moments in time.

• Variation in the number of blocking annotations. Blocking annotations indicate the distinct ways by which internet censorship was implemented. Where the implementation of internet censorship varies from ISP to ISP, we would see a high number of blocking annotations in that country. The highest number was found in Russia with 148 different annotations, while in Indonesia, we detected 84 distinct blocking annotations. In contrast, we observed fewer than 10 different blocking annotations each in Malaysia, Iran, and the UAE, while in Saudi Arabia (where ISPs consistently serve the same single blockpage) we only detected 2 blocking annotations.

• Variation in how block pages are served for LGBTIQ websites. Local ISPs in all six countries serve block pages when users try to access blocked LGBTIQ content, but the technical means by which these pages are served by ISPs vary across countries—in some cases, they even vary between ISPs within the same country. ISPs in Indonesia and Malaysia serve block pages by means of DNS hijacking, whereas Iranian ISPs serve block pages primarily by means of DNS injection. In Russia, ISPs commonly make use of HTTP transparent proxies to serve blockpages, but some Russian ISPs serve block pages by means of DNS hijacking instead. In Saudi Arabia and the UAE, ISPs deliver block pages to internet users through the use of censorship technologies.

• Some block pages in Russia contained affiliate ads. Unlike other countries, some block pages in Russia contained affiliate ads, suggesting the presence of financial incentives. We have observed ads being served as part of censorship efforts previously. In 2018, both OONI and the Citizen Lab reported on the injection of ads by ISPs in Egypt, while implementing internet censorship.

• Detection of censorship technologies in Saudi Arabia and the UAE. In both Saudi Arabia and the UAE, ISPs deliver block pages to internet users through the use of WireFilter technology, which is a network filtering device made for the ISP and commercial market, manufactured by Riyadh-based Sewar Technologies Ltd. In the UAE, we also observed blocking using a tool manufactured by Netsweeper, which is a Canadian company that sells internet filtering products to ISPs around the world.

Further details are available through our report.

We thank OONI Probe users who contributed measurements, supporting this study.

To measure the blocking of LGBTIQ websites around the world (and contribute open data), you can install the OONI Probe mobile app and tap on the LGBTIQ buttons on this page: https://ooni.org/get-involved/run/