800 likes on fake AI post: WA’s pink-slime journalism exploited by generative AI

For AI professionals, this case is a stark illustration of how generative AI can be weaponized to produce convincing, emotionally manipulative narratives at scale, posing risks to public trust and platform integrity. The ability to fabricate quotes, events, and audio — even when nonexistent — demands urgent advancements in AI detection and regulatory frameworks.

The emergence of AI-generated fake news in Western Australia, epitomized by the "Luxury Blog" and its Facebook amplifier "Voice of Aus," represents a dangerous escalation of digital disinformation. A fabricated story about shark attack victim Steven Mattaboni being on a phone call with his wife moments before the attack was spread online, complete with fabricated details about an audio recording reviewed by investigators and leaving "thousands in tears." WA Police categorically denied any involvement, and no such recording exists. Simultaneously, the same outlet falsely accused Fremantle Dockers captain Alex Pearce of shunning an LGBT-themed armband for a non-existent AFL Pride Round, with the club refusing to entertain the claims. These stories, amplified on Facebook with sensational headlines, exemplify "pink-slime journalism" — where AI scrapes and rewrites news, or outright fabricates, to generate engagement.

“

The emergence of AI-generated fake news in Western Australia, epitomized by the "Luxury Blog" and its Facebook amplifier "Voice of Aus," represents a dangerous escalation of digital disinformation.

This incident unfolds against a backdrop of rapidly advancing generative AI models capable of producing human-like text at scale. The pink-slime model has evolved from simple content aggregation to sophisticated narrative fabrication, exploiting tragedies, public figures, and emotional triggers to maximize virality. The West Australian case is particularly insidious because it preys on local communities and exploits intimate tragedies, such as a real shark attack, to create believable but entirely fictional scenarios. The use of AI to simulate audio or quoted conversations pushes the boundary of what audiences may accept as truth, especially when shared by seemingly trustworthy pages like "Voice of Aus," which cultivated 800 likes on one of its posts, as noted by Deputy Premier Rita Saffioti.

The market impact extends beyond immediate harm to individuals and institutions. Trust in digital media, already fragile, erodes further as consumers encounter these seamlessly generated fabrications. For platforms like Facebook, the spread of AI-generated content challenges content moderation systems designed for human-generated disinformation, requiring investment in AI detection tools that can distinguish synthetic text. The incident also highlights the revenue incentives behind such schemes: the fake content generates clicks and ad revenue for the operators, often with minimal cost, using automated scraping and generation tools. This economics of disinformation may incentivize more malicious actors to adopt the same techniques.