Maria Xynou, co-founder of the Open Observatory of Network Interference (OONI), has warned that internet censorship is becoming increasingly sophisticated, targeted and difficult to detect, complicating efforts to defend online freedoms worldwide. In a recent blog post, she argues that modern censorship is no longer confined to authoritarian states, but is spreading across democracies as governments adopt more advanced techniques to control digital information flows.

Democratic restrictions: According to Xynou, countries such as Russia and Kazakhstan continue to block independent media, while temporary restrictions on social-media platforms during elections or protests are becoming commonplace. Even democratic governments are experimenting with selective restrictions. Albania, for instance, blocked TikTok last year, while Spain intermittently disrupted parts of the internet by targeting Cloudflare infrastructure linked to illegal football-streaming services.

Measurement challenges: Measuring censorship, however, is far from straightforward. Technical failures, unstable networks and misconfigured domain systems can all resemble deliberate restrictions. Governments also employ a growing arsenal of methods, ranging from DNS manipulation and IP blocking to throttling internet speeds and injecting forged network responses. Because websites may be accessible on one network but blocked on another, decentralized testing has become essential.

Crowdsourced monitoring: OONI addresses this challenge through crowdsourced measurements collected from volunteers running its OONI Probe software. The organization combines local testing with open methodologies and probabilistic analysis to determine whether a service is likely being restricted. Yet uneven participation and the risks faced by contributors during politically sensitive moments, such as protests, continue to hamper comprehensive monitoring efforts.

Emerging trends: The organization’s data reveals several emerging trends. One is the rise of short-term censorship, whereby governments temporarily block platforms like WhatsApp or Facebook during elections or unrest to shape public discourse while minimizing political backlash. Another is the persistence of long-term restrictions targeting sensitive topics, including LGBTQI rights, minority issues and reproductive health. Tanzania’s sustained blocking of LGBTQI-related content is cited as one example.

Encryption is also reshaping the censorship landscape. As more websites adopt HTTPS protections, authorities are increasingly interfering directly with encrypted traffic through technologies such as Deep Packet Inspection. Rather than displaying explicit block pages, these methods often generate vague connection errors, making censorship harder for users to identify. Governments are also targeting privacy-enhancing tools themselves, including encrypted DNS services and technologies designed to conceal browsing activity.

Meanwhile, some states are experimenting with “national intranets” that limit access to approved services, effectively creating tightly controlled digital ecosystems. Iran, Russia and Myanmar are among those pursuing such allowlisting strategies.

Advocacy efforts: Xynou argues that internet measurement has become a crucial tool for human-rights advocacy. Since joining the #KeepItOn campaign in 2016, OONI’s data has supported legal challenges and advocacy efforts against shutdowns and social-media bans in countries including Pakistan, Kenya and Mozambique. With more than 3 billion measurements collected since 2012, OONI now maintains what it describes as the world’s largest open dataset on internet censorship.