We’re only beginning to tap the generative potential of artificial intelligence, and the further we push this technology, the more potent our fear of it grows. Thanks to a new report from NewsGuard (via The Verge), it looks like people are using AI to generate entire spam websites.
We have already seen people use AI to generate images, articles, books, advertising copy, videos that border on nightmare fuel, and more. There’s no doubt that it put many people out of their jobs, as it’s becoming harder to distinguish between AI content and human content. However, according to the new report, AI just took a major step forward.
People are using AI to generate spam websites
This should come as no surprise, as scores of people are using AI for their nefarious needs. Hundreds of books on Amazon are co-authored by AI, so people making entire websites using the technology won’t shock you.
The report from NewsGuard says that it has identified 49 sites. They seem to be completely written using AI. If they’re not completely written using AI, then they are almost entirely written using it. These are websites that seem to focus on posting written content like articles.
The thing about it is that, since they’re using AI, they can churn out hundreds of Articles every day. That basically gives a person the power to create a full newsroom from their bedroom.
The tells
That is scary, but there is a bit of a silver lining to this. There are obvious tells that the sites are using artificial intelligence. As sophisticated as chatbots are at creating articles, the delivery is still as clinical as a doctor’s report.
Also, since these articles are being copied and pasted right onto the website, no one is proofreading them. This means that blatant errors are being found in the articles.
According to NewsGuard, one article left an error message from the chatbot in the article. We see the text “I’m sorry, I cannot complete this prompt as it goes against OpenAI’s use case policy on generating misleading content,” appears randomly in the article. That is a dead giveaway.
Another tell is a little bit more subjective. Some of the sites have very generic names such as Biz Breaking News and Market News Reports. Ostensibly, if you’re churning out a ton of spam websites, then you won’t put that much effort into making a clever title. However, there are some legitimate sites out there whose names are a bit generic, so you won’t want to use that as your primary metric.
The best thing to do to fight off these AI spam sites is to always check for those tells. Pay attention to the wording and the phrasing of the articles and see if they sound human. Read the About pages for these sites and check out the writing staff.
Also, check out the dates for the articles. If the publication only started publishing articles a week ago, yet, it has a full writing staff of 20 people posting 500 articles every day, then it’s most likely an AI-generated site.