The headline states, “This valley is no longer safe for overnight stays — and DOC isn’t explaining why”.
The story that follows conjures a scene from a horror film. Something odd has been going on in Whakataki valley, east of Mount Aspiring National Park, the article claims.
“Today, it’s officially been closed to overnight stays — and the Department of Conservation isn’t saying exactly why.”
A former hut warden, identified only by their first name, is quoted. An anonymous DOC worker adds, “there’s more going on than we can talk about right now”.
As intriguing as it all sounds, none of it is true.
This story is among a growing number of invented stories published to harvest views at the domain morningside.nz/news. The website exists to give people information about an Auckland entertainment precinct, but the news section is full of weird, bogus stories like this one that don’t seem to belong.
There is a Whakataki, but it’s in the North Island — not in Fiordland, like the story suggests.
Another article claims DOC has put steel barriers across the entrance to Echo Hole, which it says is a cave tucked into a limestone bluff, in the South Island back country.
According to DOC, there are numerous caves in Waitaki, but none called Echo Hole. The photo used in the story appears to be generated by artificial intelligence. It shows a cave surrounded by rain forest.
“The Waitaki district in the South Island is quite a dry district. It doesn’t really have rain forest,” DOC’s manager of visitor safety and standards Andy Roberts told RNZ.
DOC staff stumbled across the stories online earlier this month. All come from the same morningside.nz website, and many follow a pattern of taking a real place name but applying it to the wrong part of New Zealand and making up a story.
In several of the stories, DOC is cast as unforthcoming about reasons for supposed closures, and often the discovery of taonga is mentioned as a possible theory. None of the stories related to DOC are true.
How a website’s news section got hijacked by AI
What is going on at morningside.nz?
Rod Ballenden, who runs the site, says it has been “hacked”. The news section was added to the site and is being populated with the articles, which are all crammed full of digital ads. They are still trying to remove the stories but it’s proving difficult, Ballenden says.
Cyber security expert Adam Boileau says he suspects the domain name may have not been renewed by the site owner. Expiry details are public, and people look out for ones soon to expire.
“They can basically snipe the domains that expire out from underneath their original owners.”
What is a little more unusual in this case is that the original website was left intact, and just a fake news section added.
“I think in this case, because it’s someone who’s trying to leverage the existing reputation. They want that value to continue. Keeping the original services and making the original site owner not notice that anything’s happened is a kind of a good way to preserve their investment,” Boileau said.
Boileau, who is technology editor at the Risky Business podcast, thinks the site owner may be able to lodge a dispute with New Zealand’s domain name registry to wrest ownership back.
“All this feels like, you know, it’s just every day on the shady modern internet. This is just what happens.”
Technology expert and author of Fake Believe: Conspiracy Theories in Aotearoa, Dylan Reeve says we may never know who is behind the hack, and whether it was a New Zealander, or someone offshore.
The likely motivation for the hacking was to make money off the advertising appearing on the stories. “It’s just filthy with ads,” he says.
The content also appears to be tailored to repeat search terms, and he notes content from it is appearing in Google AI suggestions at the top of search results.
“The fundamental thing going on here is revenue harvesting,” he says. “It’s just coherent gibberish designed to attract attention with headlines and some vaguely plausible paragraphs of content. But it’s nonsense. I would be shocked if it wasn’t AI generated,” Reeve says.
The automation may go beyond article text and image generation. There may be “automation flow” in use, Reeve says. This might include a tool which scrapes discussion website Reddit for popular topics, then uses these topics as a prompt for an AI tool to write the article and create an image.
“This can all happen completely autonomously, and you can just be pumping those out three a day, or three a minute.”
Along with numerous articles about the Department of Conservation, there are articles about petrol prices, real estate and grocery prices.
“It’s all successfully click bait-y, in a way that is sort of attached to the zeitgeist of New Zealand interests.”
Made-up comments on a made-up story
Most of the stories include some mention of cultural issues and mention iwi, a taonga, remains, and occasionally a rāhui.
Reeve thinks the choice of a divisive element to stories is intentional. “It’s probably seen that divisive topics are successful and get a lot of attention.”
In some cases, comments following the story express anger, “Totally ridiculous. As long as people are respectful to the area. Why can’t we make the most of OUR country? We pay for it. It doesn’t belong to Māori, just because they think they were there first. There was another tribe of people way before they got here”.
It’s not clear whether the comments are from real people or are fake. In some stories, the same — or remarkably similar — comments are repeated, suggesting some or all of the comments are also fake. A common pattern is for a comment starting with the word “honestly” and a reply starting with the word “exactly”.
University of Waikato research fellow Hemopereki Simon says discourse suggesting Māori aren’t the indigenous population of the country isn’t new but used to take place in books. In recent times these views have shifted online, particularly on social media.
Simon has recently studied racism in Twitter discussion on the Three Waters proposal. “When I say this stuff is out there, it’s out there,” he says.
If the comments are generated by AI, Simon is not surprised at the tone.
“AI will ultimately, to some degree, promote some type of racist tropes.”
What’s real and what’s fake
DOC’s Andy Roberts has some advice for people who stumble across fake news stories about the conservation estate.
“One of the things you’ll notice in these stories is they never quote an actual DOC person. So when DOC puts information out in a press release, there’ll always be a person that’s behind that story.”
Official notices about closures are always published on the DOC website, and Roberts urged people to visit the official site to ensure they are getting correct information, rather than rely on second-hand sources.
Closures do occur, but there is a high bar for this to happen. This could be due to a real safety risk, or occasional track or hut repairs.
rnz.co.nz
