The most effective influence operations often do not resemble traditional propaganda; instead, they may appear as maritime security seminars, expert briefs, gossip posts on Threads, pseudo-news domains, or social media accounts that seem to simply repost viral content. The true power of these operations lies not in a single dramatic falsehood but in the sophisticated infrastructure that supports them.

This evolution in influence operations can be traced through recent research into Beijing-linked cognitive warfare across the South China Sea, Taiwan’s digital landscape, and even parts of Europe. These studies invite us to look beyond simplistic notions of a centralised command, offering a nuanced view instead. The findings suggest a converging model: build legitimacy, cultivate reach, segment audiences, and maintain the machinery for timely political deployment.

To illustrate this model in detail, consider the South China Sea, where infrastructure begins at the level of expertise. A recent Cyfluence Research Center report mapped what it described as an emerging influence cluster targeting English-language debate on maritime security. The network, according to the report, combines Chinese state-affiliated institutions, researchers, and inauthentic amplification assets. The model is layered. Think tanks and research bodies supply the authoritative frame. Credentialed researchers give that frame a human face. Pseudonymous and low-credibility accounts extend their reach and create the impression of wider organic engagement.

The language used is telling. The network consistently emphasises terms such as “peace,” “stability,” “cooperation,” “prosperity,” “governance,” and a call for a “new narrative” in the South China Sea. These choices are deliberate, positioning China as a force for order and dialogue while depicting external actors, especially the United States and its partners, as sources of disruption. This is more than a dispute over reefs, coast guards, or maritime law; it is an effort to shape the very vocabulary through which these issues are perceived.

That matters for Taiwan. Beijing’s maritime messaging is part of a wider contest in the cognitive domain, in which legal claims, expert commentary, public diplomacy, and digital amplification reinforce one another. Influence here does not start with a bot. It starts with authority.

Examining a complementary aspect, the second layer of this approach centres on platform infrastructure. Doublethink Lab’s investigation into the Borderless Group, a PRC-linked content farm, is striking precisely because most of the content is not political. Researchers identified 1,217 domains hosting inauthentic news pages operated by the network, of which 475 remain active. They also attributed 549 inauthentic Threads accounts to the operation. Created between July 2023 and February 2026, these accounts produced at least 467,547 posts and replies, 83 per cent of which were designed to drive traffic to the group’s domains.

At first glance, much of this content appears banal: gossip, viral stories, plagiarised articles, entertainment posts, and engagement bait. Yet this very banality makes the threat more significant. A content farm does not need to fabricate election lies daily to become an election asset. Its primary goal is to build accounts, domains, habitual traffic, audience familiarity, and platform reach. Once this infrastructure is established, political content can be inserted whenever the situation requires.

Doublethink’s findings show signs of mature coordination: bulk account creation, proxy registration, redirect chains, automated posting, likely AI-generated summaries, copied account imagery, and repeated duplicate content.

One detail is particularly telling: 97.72 per cent of the inauthentic Threads accounts had posted duplicate content with at least one other inauthentic account. This is not the messy behaviour of ordinary users. It is an organised traffic machine.

The Borderless case also complicates the way democracies think about disinformation. The danger is not only the false post. It is dormant capacity. Taiwan’s information environment may be polluted not only by ideological propaganda but also by commercially oriented networks that have already built the distribution rails through which political material can later move.

These infrastructure patterns raise the stakes for upcoming political events. The likely stress test is Taiwan’s November 2026 local elections. A Global Taiwan Institute analysis reports that at the CCP’s annual Taiwan Work Conference in February 2026, Chinese officials discussed establishing a task force to interfere in those elections, including through united front work in cyberspace aimed at damaging “Taiwan independence forces.”

The same report notes that Taiwan’s National Security Bureau recorded a 60 per cent increase in inauthentic social media accounts between 2024 and 2025, and more than 2 million instances of disinformation in 2025 alone.

Local elections seem less geopolitically charged than presidential polls, but this apparent neutrality can make them more vulnerable to manipulation. Their fragmentation and local specificity mean grievances over housing, agriculture, tourism, corruption, patronage, infrastructure, food prices, and candidate reputations dominate—exactly the types of issues information operations can exploit. Beijing’s goal is not to change every vote, but to shift the environment in which voters assess trust, risk, competence, and political fatigue. That makes local elections attractive targets for influence operations: they are issue-driven, locally fragmented, and often less visible to national monitors.

There is also a strategic timeline. Taiwan’s 2026 local races will help shape party morale, candidate pipelines, and political momentum ahead of the 2028 presidential election. Local power in Taiwan is not merely local. It feeds national politics, shapes cross-Strait narratives, and creates openings for economic inducement, elite cultivation, and social division.

The GoLaxy leaks provide an additional layer of understanding of where influence tactics are heading. The leaked material, as discussed by researchers, exposed a PRC-based “influence-for-hire” firm that used AI-powered systems to manipulate and interfere in foreign information. One project reportedly collected 50,000 Taiwan-related news items and classified political actors into categories such as “Hardliners,” “Moderates,” “Swing Voters,” and “Objectivists,” with at least 1,000 key individuals in each category.

This marks a shift from mass persuasion to precision influence.

The future influence operation will not necessarily broadcast one crude message to millions. It may classify audiences, track political moods, identify emotional vulnerabilities, generate tailored content, and deploy it through accounts that appear realistic enough to escape casual detection.

The target will not be limited to opinion, but trust.

The same logic is visible beyond the Indo-Pacific. A NATO Strategic Communications Centre of Excellence report on China’s influence in Denmark and Lithuania identifies multiple avenues through which Beijing seeks to shape political environments, including public communication and diplomatic pressure, academic ties, economic exposure, infrastructure, United Front networks, and espionage. Lithuania reflects the punitive model: after Vilnius allowed the opening of the Taiwanese Representative Office, Beijing responded with diplomatic downgrading, economic pressure, and information pressure. Denmark, by contrast, reflects a quieter model of elite access, backstage pressure, and what the report calls “Shadow Wolf Warrior” diplomacy.

The comparison matters because it shows that Beijing’s influence architecture adapts to the theatre. In the South China Sea, it can look like maritime expertise and pseudo-independent amplification. On Threads, it can look like gossip and ad traffic. In Taiwan, it can become election-focused cognitive warfare. In Europe, it can take the form of coercion, elite access, or backstage intimidation.

For Taiwan, the lesson is not that every falsehood can be corrected in time. Fact-checking remains essential, but it is a late-stage intervention. By the time a rumour, forged document, deepfake, or manipulated narrative has entered public circulation, the first line of defence has often been breached. Taiwan, therefore, needs to move upstream: from rebutting individual claims to detecting the systems that enable manipulation to scale.

That begins with mapping infrastructure, not just content. Researchers, election authorities, civil society groups, and relevant agencies should systematically track suspicious domains, fake accounts, content farms, ad-tech incentives, redirect chains, AI-generated content pipelines, and repeated narrative patterns across platforms. The question should not only be whether a post is false. It should be whether a network is being prepared for future political activation.

This is especially important before the 2026 local elections. Local races are vulnerable not because they are less important, but because they are more fragmented. A presidential election attracts national scrutiny. A county-level rumour about agricultural imports, a mayoral scandal, a housing grievance, or a manipulated story about local patronage may travel further before it is noticed. Taiwan should therefore establish a standing election information-integrity mechanism linking the Central Election Commission, relevant agencies, independent researchers, fact-checkers, local media, civil society monitors, and platform contacts. Its role should not be to police legitimate political debate, but to detect coordinated inauthentic behaviour, foreign-linked amplification, and suspicious domain networks before they reach scale.

Taiwan should also invest in technique-based prebunking. Before the 2026 local elections, voters should be exposed to the playbook: forged documents, fake local outrage, impersonation, emotional baiting, false dilemmas, synthetic consensus, and coordinated amplification. This should be done at the local level, not only through national messaging. A farming county, a technology hub, a port city, and a tourism-dependent district will not be targeted through identical narratives. Prebunking should therefore be tailored to local communities' vulnerabilities while remaining politically neutral. The goal is not to tell citizens what to believe, but to help them recognise when public opinion is being engineered.

Platform accountability is the next priority. Taiwan cannot defend an information space that it is not allowed to see. Companies operating services such as Facebook, Threads, TikTok, YouTube, and LINE should be pressed to provide faster election-period reporting on account-creation spikes, coordinated sharing of suspicious domains, foreign-origin activity, political advertising, and repeated amplification by low-credibility networks. The goal is not to impose censorship but to enhance transparency. Platforms are no longer neutral venues during election cycles; they are part of the democratic security environment.

Taiwan must also prepare for the AI turn. The GoLaxy material suggests that future influence operations may classify political actors, identify audience segments, generate tailored narratives, and deploy them through realistic digital personas. That requires rapid authentication channels for candidates, journalists, public officials, and local government offices targeted by forged documents, synthetic audio, or deepfake video. It also requires public databases of known manipulation tactics and pre-election AI red-teaming exercises that test how quickly false or synthetic material could move through Taiwan’s information ecosystem.

As AI systems increasingly ingest and summarise public information, Taiwan should also prepare for machine-facing influence. Official clarifications, election notices, fact-checks, and emergency communications should be machine-readable, provenance-rich, and easy to verify. In the AI era, the public record must be written not only for citizens, but also for the systems that will summarise it.

Yet the most important defence is trust. Beijing’s influence operations are often designed less to persuade citizens to love China than to make them doubt their institutions, leaders, media, allies, and one another. When authorities communicate slowly or defensively, information gaps become fertile ground for manipulation. During cyber incidents, platform disruptions, forged-document campaigns, or sudden waves of rumours, Taiwan’s public agencies should respond quickly, clearly, and with as much evidence as possible. The aim should not be to tell citizens what to think, but to prevent uncertainty from being exploited.

The legitimacy of Taiwan’s response will depend on democratic guardrails: political neutrality, public oversight, no domestic political targeting, and strictly defensive use of automation. Taiwan does not need a Ministry of Truth. It needs a trusted bridge between election authorities, researchers, platforms, civil society, and local media.

China’s influence machine now resembles a grid more than a megaphone, moving narratives, incentives, pressure, and artificial consensus across interlocking channels. Taiwan’s task is to detect that grid before it is fully mobilised. Its best defence will not be panic or censorship, but transparency, technical preparedness, platform accountability, and a democratic public that understands how influence infrastructure works before the next election cycle tests it.

(This piece reflects the author's opinion, and does not represent the opinion of CommonWealth Magazine.)

CommonWealth Magazine welcomes op-ed submissions. Please send your article proposals to [email protected]

About the author:

Aritra Banerjee is an AI content evaluation and strategic communications specialist at a U.S.-based AI and robotics firm, where he evaluates AI-generated content for accuracy, neutrality, and compliance in security and geopolitics-related domains.

Have you read?

• Winning the Cognitive Fight: How Taiwan Can Turn Information Defence into Strategic Communication
• Fake Press, Real Spies: How China Tried to Infiltrate Taiwan’s Civil Society

Uploaded by Ian Huang