Canhsat4sao

Private gatekeepers: how ISPs become instruments of censorship

When governments lack the technical capacity or political appetite to censor the internet directly, they frequently delegate that authority to private telecommunications companies. The result is a quiet, distributed form of information control that rarely makes headlines but shapes what millions of people can read, watch, and share. Internet service providers, the commercial entities that connect households and businesses to the global network, become de facto gatekeepers whose compliance decisions determine the boundaries of public discourse.

This phenomenon, sometimes described as censorship by proxy, has been a defining feature of Vietnam's digital governance strategy throughout the past decade. Yet the model extends well beyond Hanoi, and Australian readers will recognise elements of it in the regulatory frameworks that govern their own internet infrastructure.

Delegated authority and the compliance economy

The modern state does not need to build its own filtering apparatus when it can compel private firms to do the work. Through licensing requirements, administrative directives, and informal pressure, governments can transform commercial ISPs into extensions of the censorship apparatus without ever passing a single blocking law through parliament.

In Vietnam, this arrangement has been formalised through a series of decrees issued between 2010 and 2016 that required domestic telecom companies to remove content deemed threatening to national security, public order, or social morality. Companies that failed to comply risked having their operating licences revoked, a threat that proved more persuasive than any criminal prosecution. The compliance economy that emerged rewarded firms that pre-emptively restricted access to politically sensitive material, blogs critical of ruling officials, and diaspora news outlets reporting on human rights cases.

The incentive structure mirrors what scholars have observed in other jurisdictions. When the cost of compliance is lower than the cost of resistance, rational commercial actors choose to filter, even when doing so conflicts with their stated corporate values.

Australia's regulatory precedents

Australian audiences may find this pattern familiar, even if the political stakes differ. The Telecommunications (Interception and Access) Act 1979, amended repeatedly over the past two decades, grants law enforcement agencies extensive powers to compel carriers such as Telstra, Optus, and TPG to retain and surrender customer metadata. While metadata retention is not the same as content blocking, it establishes a legal tradition of treating ISPs as instruments of state policy.

More directly relevant is the censorship regime overseen by the eSafety Commissioner and the Australian Communications and Media Authority. The list of blocked URLs maintained under the Broadcasting Services Act includes material related to child exploitation but has occasionally expanded into adjacent categories. The mechanism works precisely because ISPs implement the blocks at the network level, rather than requiring individual users to install filtering software on their devices.

In Melbourne and Sydney, where the vast majority of Australian internet traffic originates and terminates, this infrastructure operates invisibly. A resident attempting to access a blocked overseas blog would simply receive a timeout error, with no indication that a private company had decided the content was unsuitable.

Mechanisms used at the network level

The technical methods used by ISPs to enforce censorship are varied and increasingly sophisticated. Several approaches have become standard across filtering states:

Each technique carries different costs and precision. DNS poisoning is cheap but easily circumvented. Deep packet inspection is more thorough but expensive to maintain and raises privacy concerns. The choice of method reflects the political priorities of the regulating authority rather than purely technical considerations.

The cost to legitimate speech

Filtering regimes rarely achieve surgical precision. When an ISP is ordered to block a specific blog or news outlet, the technical implementation often affects thousands of unrelated websites hosted on the same server or sharing the same IP address. This collateral damage has been well documented in Vietnam, where small businesses, academic researchers, and diaspora community groups have found their websites inaccessible because of politically motivated blocking orders directed at completely different content.

The chilling effect extends beyond the directly targeted. Writers, journalists, and ordinary citizens who observe that certain topics or sources are unreachable often conclude that engaging with similar subjects carries risks they wish to avoid. Self-censorship becomes a rational response to an opaque filtering environment, and the diversity of public discourse contracts accordingly.

Resistance and structural limits

Vietnamese civil society has responded with a mixture of technical adaptation and legal advocacy. VPN usage surged during periods of heightened political tension, and diaspora media outlets experimented with mirror sites and decentralised publishing platforms. Lawyers representing detained bloggers attempted to use domestic courts to challenge the legality of specific blocking orders, though such challenges rarely succeeded.

The deeper problem is structural. When the entities responsible for filtering are also the only entities with the technical capacity to provide internet access, citizens have no commercial alternative. Choosing a different ISP means choosing the same filtering regime, because the regulatory requirement applies across the industry. Competition does not produce diversity of practice when the underlying rules are uniform.

Signals of delegated filtering

Readers and researchers can identify censorship by proxy through several observable signals in any given jurisdiction:

These indicators rarely appear in isolation. A filtering regime typically combines several of them, creating a system in which private firms absorb the legal and reputational costs of censorship while the state retains deniability.

The uncomfortable truth is that most of the world's internet censorship is performed by companies that would prefer to remain neutral. ISPs are not ideological actors by nature; they are utilities seeking predictable regulatory environments. When governments make censorship a condition of operating, the firms comply because the alternative is bankruptcy or criminal sanction.

Australian readers who value the relatively open character of their domestic internet should remember that this openness is contingent on political choices, not technological inevitability. The same private networks that deliver streaming sport from Melbourne stadiums and banking services from Sydney headquarters can, with a change in law or administrative practice, be redirected toward suppressing dissent. The infrastructure of control is already in place; what changes is whether it is activated.