A few weeks back, I was looking into possibly getting a PlayStation 5, and I was researching whether I could use my Xbox controller on my hypothetical PS5. I searched this, and was served the following initial result:
Initially, that's a pretty emphatic conformation, right? Except, these two articles that popped to the top are actually saying very different things. Robots.net asserts that all you have to do is connect the controller; it's that easy! Meanwhile, Alvaro Trigo is saying that the PS5 doesn't natively support it, but there are some obscure workarounds using adapters and other third-party technologies to trick the PlayStation into cooperating. What's happening here?
I can't know for certain, but robots.net just reeks of artificial intelligence. More specifically, AI hallucination, where chatbots like ChatGPT confidently assert complete nonsense. We are not yet at a point here we can determine for certain that something has been written by AI; if you see a program or service that claims the capability to do that, they're selling something. However, there are certainly clues.
I believe that this robots.net website is at least partially, likely entirely, AI generated. It took me less than five minutes to discover the truth about connecting my Xbox controller to a PS5; I can't. At least, not practically. Any journalist worth their salt, being paid to write an article for a website, should have figured that out. They certainly would not have confidently asserted blatantly false information like they have done on robots.net.
So, I dug deeper into robots.net. Who is Anthiathia Vail, the supposed author of this article? Helpfully, her name links to an about the author page. Unhelpfully, that page looks like this:
Searching online for Anthiathia Vail brings back no results besides her articles on robots.net, except for a journalist aggregation website called Muck Rack...which has no information besides her articles on robots.net. What's more, the picture is too low resolution to reverse image search.
A cursory look at the other "authors" on robots.net brings back similar results; either obscure names that somehow lack any other web presence, or names so generic that they bring back dozens of people, all paired with low resolution photos that could very easily be AI-generated themselves. So, I can now say with near absolute certainty that robots.net is a website created almost entirely by artificial intelligence, but why should you care?
In the grand scheme of the internet, one AI-generated website may seem pretty meaningless. However, robots.net is just a small part of a much grander crisis on the internet now, and it is one of the most mundane and relatively harmless offenders. Robots.net is scummy and dishonest, but it is ultimately more a scheme in search of advertising revenue than anything else. Other websites using robots.net's methods can be significantly more malicious.
The fact that a website that was (allegedly) entirely AI generated popped to the top of a simple search query is the real danger at play here. Search Engine Optimization, or SEO, is the process of, well, optimizing a website for search engines like Google and Bing. Clearly, whoever is behind robots.net is very good at SEO, and they are not alone. Malicious actors are working to create AI-generated webpages with excellent SEO for the purpose of easily and efficiently spreading malware to victims all over the globe.
Many of the world's most popular websites might lack the greatest SEO, not for lack of trying, but because the rules of SEO are always shifting. This means that there is a vulnerability of sorts where bad actors can spoof popular websites, and even if only temporarily, beat out the actual site. To use one of the most commonly spoofed sites as an example, searching for the US Postal Service may return a website like uspostal.com first (do NOT go to that link; it was dead when I wrote this, but that might have changed), which is obviously not the actual USPS website (it's usps.com), but how often do you read the web address word for word, letter by letter when you search something?
These schemes take advantage of people's trust and/or ignorance to bring them to pages filled to the brim with malware. Ideally, these sites would be blocked or flagged by your device's security software, but that is naturally dependent on your device having such software. Regardless, the safest option is knowing not to go to these sites in the first place. With that in mind, consider the following best practices when using the internet in this age of rampant AI use:
Try and stick to websites that you know and trust already.
Go to those websites, using their URLs (for example, go to burkitc.com, don't just Google "Burk I.T.").
As always, avoid clicking suspicious links, even if the link is on a site, document, or email that you initially trust.
Use security software like MDR, EDR, and XDR to protect your device (for more information on what specific security solution is best for you, contact us).
Website spoofing has been going on for years, but generative AI has made it faster, easier, and more convincing than ever before. It is just another in a long line of crimes and other threats made worse by artificial intelligence. It's not all bad; AI in many ways is just another new tool, and it can be used for good as well. However, the ease of use for AI tools like ChatGPT have made scammers more efficient and effective than ever before, and it never hurts to have an abundance of caution online.
Comments