AI-generated fake news pages on Facebook scraped legitimate New Zealand news stories, rewrote them with AI, and paired them with unlabeled synthetic images that dramatized events and manipulated photos of real people, including deceased individuals.
At least 10 Facebook pages were identified taking existing New Zealand news stories, running them through artificial intelligence to rewrite them, and publishing them with synthetic images. One page called 'NZ News Hub' with over 4,700 followers was analyzed, showing 209 posts in January 2026 with no original reporting. None of the AI-generated images were labeled as synthetic, and some featured fabricated photos of real people, including a manipulated image of a minor killed in the Mount Maunganui landslide shown dancing, and edited photos of parents who lost their daughter to suicide. Natural disasters were consistently dramatized beyond reality, with authentic highway slips depicted as more destructive, and a grounded tourist boat edited to appear packed with more passengers. Police were often shown in incorrect British or American uniforms with guns drawn when no indication existed that officers were armed. Some posts accidentally included raw AI prompts like 'Here's a news-style rewrite with a clear headline, emojis, and top hashtags.' Individual posts regularly attracted over 1,000 likes and comments, with many users criticizing the fake images and blaming 'the media' for fake news, though these pages had no connection to legitimate news organizations. Google Image searches revealed several pictures contained 'SynthID' digital watermarks indicating they were created using Google's AI image-generation tools.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
AI system
Due to a decision or action made by an AI system
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
