- Filed
- September 10, 2026, 17:00 UTC
- Actor
- Sri Lanka-based influencer/content business operating commercial AI-slop networks (Reset Tech attribution; parallel activity in multiple countries)
- Target
- Australian public political discourse and voters; secondary audiences in other Western democracies via parallel pages
- Confidence
- high Multiple independent open sources, low disagreement about the facts
- Techniques
- Synthetic mediaAstroturfingCard stackingFlooding the zone
- Channels
- Facebook pages and Reels using AI deepfakes of real politicians
- lure content (animals, beauty, patriotism) transitioning to political bait
- platform monetization payments and engagement farming
- identical branding and coordination across hundreds of pages
- cross-country reuse of the same operator infrastructure
Summary
Reset Tech research published via ABC News on 9 September 2026 documents a coordinated network of more than 200 Facebook pages targeting Australian political figures with AI-generated deepfake Reels and images. Pages impersonate or feature Anthony Albanese, Angus Taylor, Pauline Hanson, Fatima Payman, David Pocock and others in fabricated inflammatory statements on migration, Islam and national identity. Approximately 10 percent of pages are monetized under Meta’s creator program. The same Sri Lanka-linked operators run parallel networks across France, Italy, UK, US, Canada and Mexico, totaling 1,864 pages and 5.9 million followers. Meta stated it is reviewing the assets.
Analysis
Primary sourcing is Reset Tech’s investigation, reported by ABC News on 9 September 2026 with contemporaneous confirmation from targeted politicians and a Meta review statement. Researchers tracked more than 200 Australia-focused pages over eight months, identifying coordinated branding, posting patterns and links to a Sri Lankan influencer business. Content consists predominantly of short AI-generated video clips in which real Australian politicians appear to deliver inflammatory statements they never made—anti-migrant, anti-Muslim or nationalistic themes, plus sexualized imagery of female politicians. Pages often open with non-political lure material (AI animals, scenic Australia) before pivoting. Roughly one in ten pages receives direct Meta creator payments. The identical operator cluster maintains 1,864 pages globally with 5.9 million followers, applying the same template in multiple countries.
Mechanism: synthetic-media supplies low-cost, high-volume deepfakes that exploit the human tendency to treat video of a known face as authentic speech. Astroturf is achieved by presenting the pages as organic Australian community or news surfaces rather than foreign commercial operations. Card-stacking appears in the selective, consistently inflammatory framing that privileges outrage-inducing narratives while omitting context or opposing evidence. Flooding-the-zone is the industrial scale—hundreds of pages, continuous production, A/B testing via engagement metrics—to saturate feeds and maximize algorithmic distribution. Platform monetization converts attention into revenue, creating a self-reinforcing commercial incentive independent of any state directive.
This qualifies as an influence operation because origin is concealed, coordination across assets is demonstrated, and the activity systematically distorts the information environment around real political actors and contested issues. Commercial motivation does not remove the propaganda character; the techniques remain identical whether the goal is profit, ideology or both. Technique scoring is actor-agnostic: the same combination of synthetic personas/media, false local authenticity and volume flooding has been documented in state-linked and purely commercial settings. Limited platform labeling (one in three posts) and low removal rates amplify residual exposure. Meta’s post-publication review raises detectability but does not erase prior circulation.
Literacy counter: treat short-form video of politicians delivering unusually inflammatory or out-of-character statements—especially on pages with generic patriotic or lifestyle branding—as provisional until verified against primary footage or official channels. Check for AI artifacts, missing source attribution and rapid follower or engagement patterns. Prefer named, accountable media or direct politician channels. Platforms should enforce consistent synthetic-media labeling, restrict monetization of coordinated inauthentic political content, and publish behavioral fingerprints of commercial deepfake clusters so independent researchers can surface them earlier.