Influence Campaigns in the Philippines: 2025 Midterm Elections

A sketch on a background image, mostly in turqoise color and with various shapes
The Philippines was dubbed “patient zero” of the post-truth era just shy of a decade ago. His presidency normalised malign influence as a tool of the state, with his top propagandists appointed to government positions. In turn, players across the political aisle set up their own operations.

The Philippines was dubbed “patient zero” of the post-truth era just shy of a decade ago. Then-president Rodrigo Duterte, a populist known for his war on drugs that killed thousands, sailed to office with the aid of a social media campaign that Oxford Internet Institute estimated to cost $200,000. His presidency normalised malign influence as a tool of the state, with his top propagandists appointed to government positions. In turn, players across the political aisle set up their own operations. By the next presidential election in 2022, researchers estimated the cottage industry of political influence had grown to 1.5 billion Philippine pesos, or some €45.6 million.

Today, pro-Duterte influence is a cyborg that is part machinery, part organic support. Like in previous elections, disinformation operations attempt to project an illusion of majority via astroturfing and copypasta schemes.

Astroturfing is the practice of hiding the origins of an orchestrated message by making it appear as though it originates from (and is supported by) unsolicited grassroots participants. Copypasta in the context of influence campaigns refers to when posts perpetuate false information by copying and pasting text and passing them along to others.

To soft supporters and undecided voters, their narratives are designed to evoke a bandwagon effect, provoke heightened emotions and sympathy, or harass dissenters into submission. But even if they do not succeed at converting views to votes, they have already achieved their goal by sowing chaos and confusion, making it more discouraging to participate in public life.

This article explores some of the ways that influence has developed between the 2022 election and the recent 2025 midterms. Structures of political influence remained largely similar to that of previous elections: influencer pages cascaded a message, and it was widely engaged with and spread, both by troll accounts and real supporters. In 2019, a troll farm employee told me social media platforms were cracking down on such activity, pushing them to camouflage inauthentic accounts better by cultivating them to post non-political content, as well as personally customising their messages. In 2022, micro-influencers became popular with the rise of short-form video and the search for authenticity. But if influence operations of recent elections shifted towards “smart” trolling and on-camera influencers, recent trends point to a backslide towards more generic and basic campaigns.

These trends may relate to the overlap between international politics and technology firms: following Donald Trump’s return to the U.S. presidency in 2024, Silicon Valley’s disinterest in resolving the spread of disinformation on social media is now out in the open. Meta CEO Mark Zuckerberg announced he would stop supporting fact-checking initiatives, which has worried partners in the Global South. The closure of transparency tool CrowdTangle has impacted local researchers. Disinformation research also took a hit following cuts in U.S. funding. Among those affected is nonprofit Internews’ Manila office, which supported studies in the field and released a monthly newsletter tracking influence operations.

On the ground, this allows disinformation actors to return to old tricks. In the last year, rampant astro-turfing and copypasta schemes indicate that Facebook’s previously slow — but at least present trajectory towards catching coordinated inauthentic behaviour — has stalled. These tactics are now aided by AI, which has accelerated disinformation production through deepfakes and automating customised text. Despite these challenges, investigations and research have continued over the last year.

Influence shaped by political polarisation and feuds

The midterm in May this year was defined by the split of two ruling political dynasties: The Marcoses and the Dutertes. In 2022, Ferdinand “Bongbong” Marcos Jr., son of the late dictator, succeeded Duterte. The latter’s daughter, Sara Duterte, was his vice president. The Marcoses were themselves masters of political spin, benefitting from conspiracy theory videos and TikTok fancams that whitewashed their father’s bloody and corrupt regime.

However, the Marcos-Duterte alliance has since soured. Among the first indicators was a split among the influencers who propped them up, as they failed to organise into a group to gain access to presidential coverage. The fallout coincided with, and involved disagreements about, the departure of Marcos’ then-executive secretary Vic Rodriguez. Rodriguez would later go on to run for the Senate under Duterte’s banner.

The feud between the two families escalated when Sara Duterte threatened on a livestream to assassinate the president, following a House investigation into millions of pesos worth of anomalous receipts issued by her office. In March, the Marcos government conducted a high-profile arrest of the Duterte patriarch - sending him to face charges of crimes against humanity at the International Criminal Court.

In the May midterms, twelve Senate seats were split evenly between Marcos’ administration and the Duterte-endorsed slate, with five on either side. The last two seats went to a surprising comeback for two senators under the liberal opposition, critical of both Marcos and Duterte. Arguably, the highest stake of the consequent election results is the anticipated impeachment court trial of Sara Duterte on charges of her assassination threat and corruption scandal. A conviction would bar her from running for the presidency in 2028.

Ads, smear and inauthentic behaviour on Meta

Some advertising campaigns can be tracked through transparency mechanisms. Data analysed from Who Targets Me, a nonprofit advocating for political ad transparency and accountability, shows that in the last three months leading up to the 2025 election, the highest spenders on Facebooks ads were: Camille Villar, a political dynast from the richest family in the Philippines, who secured endorsements from both the Marcos and Duterte camps, and spent a total of PhP 25 million (est. €337,000); liberal opposition bets Bam Aquino and Kiko Pangilinan, across five different Facebook pages that included volunteer-led initiatives; and Bong Go, a former Duterte assistant who topped the Senate race.

However, ad library data only accounts for above-board spending, and a lot of the campaigning techniques used in the last decade in The Phillippines has been more obscured. For example, in 2022, liberal opposition candidate Leni Robredo was the highest spender on Facebook ads; Marcos, who would come to win the presidency, registered none — but inauthentic behaviour in his favour was widespread across various platforms. In another case, ads which showed the use of misleading information or obscured identities were masquerading as media outlets and “collectively paid over PhP 48 million on political advertising,” according to a Philippine Center for Investigative Journalism report this year.

Following the election, smear operations targeting Aquino and Pangilinan — who finished second and fifth in the Senate race — surfaced. For example, a widely popular free college tuition law came on the receiving end of a copypasta attack crediting Duterte instead of Aquino, its original author. Pangilinan, meanwhile, was mocked by accounts with Vietnamese names for the way he ate soup on a food blog. He called the attacks “orchestrated” and “well-funded”, calling on election authorities and Meta to act. After the election, Pangilinan was also misleadingly discredited for a law he authored on juvenile justice in another series of copy-pasted posts.

The Commission on Elections prohibited deepfakes and “false amplifiers” like bots and trolls on the campaign for the first time this year. It also required candidates to fully disclose their campaign spending. However, politicians, as has been shown, can often skirt the rules, such as by conducting influence operations through undeclared fronts.

AI-aided Duterte vs Marcos feud

Duterte’s arrest at the ICC just two months before the election sparked an onslaught of disinformation. According to monitoring by fact-checker Vera Files, over 80% of narratives favoured Duterte. The dominant framing cast Duterte as a victim. Supporters called his arrest a kidnapping — a phrase popularised by his daughter on Instagram on the night of the arrest. YouTube influencers cashed in on the pro-Duterte algorithm, earning up to an estimated PhP 20,000 a day, according to calculations by Rappler.

Among the fake news feeding into this complex were repurposed photographs of protests abroad and other mass gatherings as supposed rallies for Duterte, and false allegations of mass resignations among the military and police. Vera Files traced the latter to a handful of misleading accounts with a formulaic script that indicated AI use. But perhaps the most troubling of these narratives is the denialism of extrajudicial killings, resulting in widespread harassment of drug war victims’ families.

Following the election, TikTok revealed that it took down a million accounts and three networks for inauthentic behaviour in the two months leading up to the election. It did not reveal the beneficiaries of such operations. Earlier this June, OpenAI revealed its tool ChatGPT had been used to analyse content and generate comments and promotional material for a pro-Marcos influence campaign targeting the Dutertes. OpenAI found the content used on Facebook and TikTok, and subsequently banned its users, which it traced to a public relations company called Comm&Sense Inc. Its social media pages and website are now offline.

While interventions to remove accounts are not unwelcome, the opaque and selective nature of such platform takedowns appear to be a mere chip in the larger disinformation machine, which churns on. On TikTok, the Dutertes have seized a previously pro-Marcos algorithm to their favor. Its content includes an influx of pro-Duterte AI-generated imagery, from non-existent sculptures and monuments to generative Duterte baby photos. While takedowns are typically centered around elections, much of this harmful content is only election-adjacent, packaged as entertainment, and are seeded year-round. They also do not fit into the mold of AI-generated political endorsements, which are against platform rules, and therefore skirt removal.

Digital Influence and Red Flags from China

Recent trends point to pro-Duterte influencers representing China’s interests in their content too. Beijing “increasingly [relies] on indirect messaging, using intermediaries to shape or convey its narratives without direct attribution,” research lab AidData said in a September 2024 report. A paper released by the think tank Stratbase Institute lists some narrative devices deployed by pro-Duterte bloggers, including inflating their own credentials. Philstar.com found from a sample of 30 videos that six different pro-Duterte vloggers tended to promote Beijing’s point of view through distraction, blaming the United States, and fear-mongering or exacerbating a full-blown war. Accounts traced to China have also had a hand at sowing distrust in the Marcos administration, particularly amplifying Duterte’s call for the secession of the southern Philippines, and spreading a deepfake of Marcos allegedly using drugs. The latter video was also spread by pro-Duterte influencers.

Coming under scrutiny at the House of Representatives, pro-Duterte bloggers admitted in March that a number of them attended a state-sponsored content training seminar in Beijing. Senator Francis Tolentino also accused the China Embassy of hiring a Metro Manila-based firm InfinitUs Marketing Solutions Inc., for troll services. The company denied the allegations.

It has yet to be seen whether Beijing’s efforts successfully make a dent on a public that is distrustful of China. However, anti-Marcos messaging might resonate with viewers, and the package deal of pro-Duterte and China propaganda can make the latter more palatable to a passive audience. Across platforms, partisan propagandists who are disinformation spreaders and serial violators of community guidelines continue to enjoy large followings.

Recommendations: Nuanced Problems and Comprehensive Solutions

On the surface, polls show that public opinion is hot and cold with the Dutertes. Surveys indicate Filipinos favour an ICC trial for Duterte - although Pulse Asia found a recent shift, potentially due to disinformation. Sara Duterte enjoys higher trust ratings than Marcos, but nine in ten Filipinos believe she should address the charges against her. These seemingly contradictory results suggest that winning public opinion is more nuanced — and therefore, more competitive — than social media discourse lets on. More Filipinos are likely to have more malleable and negotiable opinions depending on the evidence and information they receive.

Online influence operations are also an extension of offline infrastructure and practices. In the 2025 midterms, local government endorsements and policy-based campaigning were also factors in the success of some candidates. Any effective counter-disinformation approach must similarly engage audiences both offline and online. One overwhelming problem is the asymmetry in the quantity of pro-Duterte malign content versus quality fact-checked content. In the online realm, one solution would be to increase the supply of counter-narratives to Duterte, as cultural critic Katrina Stuart Santiago points out. Even if such content does not go as viral as harmful content today, the presence of SEO-hit quality information must be seeded to pay off both in the short term, for people looking for clarifying analysis on ongoing issues, and the long term, to prevent extremely partisan default algorithms in future elections. At the same time, there should be continued reporting and pushing for takedowns of malign actors and disinformation superspreaders. The suspension of SMNI, a media outlet owned by doomsday pastor and Duterte ally Apollo Quiboloy, on Meta and YouTube is one example of a repeat platform offender being held to account.

Finally, democracy advocates must push for various solutions offline: community-level depolarisation initiatives, media literacy and disinformation resiliency, platform accountability, and continuous support for academics and journalists to continue documenting, exposing, and researching malign influence and its appeal. Any public policy addressing disinformation should consider campaign finance and platform transparency, data privacy, AI ethics, cyber security, and information integrity over broad strokes, vaguely worded censorship laws.

Author: Regine Cabato

Regine Cabato is a journalist. She was previously a Manila-based reporter for the Washington Post for over five years. She was a fellow at the Reuters Institute for the Study of Journalism, where her project covered newsroom responses to influence operations. She has completed training programs by the Global Investigative Journalism Network and Tactical Tech on digital investigations and the influence industry. She is currently pursuing a master's degree in politics and international relations at SOAS University of London.