Don’t Click the GTA 6 Demo, Don’t Trust the Flirty DM: This Week’s Cybersecurity Survival Guide | #hacking | #cybersecurity | #infosec | #comptia | #pentest | #ransomware

The Grand Theft Auto VI leaks keep coming. Just this week, additional footage has been leaked, Rockstar Games fired off subpoenas and legal threats to try to find the hacker, and the hacking group itself even started asking for money just to talk about it. Rockstar finally issued a statement on the leaks, apologizing for the long wait for the game, which left a hole that leaks would naturally fill. It claims to have no plans to change its marketing or launch strategy. Meanwhile, the latest GTA VI-related scam promises you a demo of the game before launch. Don’t fall for it, it’s malware.  

Speaking of things that seem too good to be true, this week we also covered the explosion of romance scams and how to protect yourself from them. In an age when AI makes scamming easier than ever, I hate to break it to you, but that lovely guy, gal, or non-binary pal that slid into your DMs to flirt with you may not be a real person, and they may actually just want to get you to open your wallet and hand over your money or personal information. 

Additionally, principal security writer, Neil Rubenking, who has reviewed virtually every antivirus program and security suite on the market, took a look at Splashtop Shield, a new antivirus from the company that makes excellent remote access and management tools. Unfortunately, the antivirus side of the app isn’t quite there yet, but we’re excited to see where it goes. More companies that focus on other services are starting to bundle antivirus software with their products, which we love to see, as long as they actually do a good job at protecting you. 

Let’s check out what else has been going on in the infosec world this week.

‘Near-Autonomous’ AI Agents Attack Taiwan’s Nuclear Safety Agency

If there’s any news that’s been lighting up my inbox lately, it’s about how more and more people are reporting hacks by near-autonomous AI agents. In this case, The Register reports that suspected Chinese hackers used publicly available AI tools to attack Taiwan’s nuclear safety agency, its supply chain vendors, and several associated energy companies. Security researchers investigating the attack called it “near-autonomous,” with activity that seemed entirely automated. That didn’t seem to diminish the attack’s efficacy, though, as over 2,500 personnel records were extracted via more than 85 compromised accounts. 

While the target of the attack is certainly important (and terrifying), the most noteworthy aspect is that it was carried out with minimal human intervention and using AI tools anyone can access. While it’s important to remember that autonomous AI attacks have gone from conceptual to common in a matter of months, AI is still just a tool and doesn’t bear culpability. The people who use it, develop it, and maintain it are the ones responsible for preventing abuse like this. 

Fake Microsoft Security Scans Trick Victims Into Uninstalling Their Antivirus

You would think it goes without saying that if some app or website tells you to uninstall your antivirus software, you shouldn’t. But what if you stumble on a website that looks like it’s a Microsoft official one, complete with a very official-looking video with a man in a suit explaining that your antivirus has been causing system problems, that you’re due for a refund for your subscription, and that they’ll contact you in order to process it once you’re uninstalled? Yup: It’s still a scam, albeit a very well-produced one. 

We learned about the scam over at the Malwarebytes Blog, where the team has observed nearly a dozen of these scam sites on the same host, all of them calling themselves something like “SysScan” or something designed to sound like Microsoft. The full blog post breaks down the scam in detail, including how both the video and the website itself show signs of being AI-generated, along with a point-by-point explanation of how these kinds of scams work. It’s worth checking out. 

Here’s how the scam works: You stumble on the website, which says your PC has system problems (that part is probably familiar), but to fix it, you need to uninstall your antivirus software and provide your name and phone number so they can contact you to process a refund for your subscription. They even play a lookalike AI-generated video asking you to hold on and saying someone will contact you in three to five minutes. Then, surprisingly, they actually do call you, asking for more personal information to “verify” your identity, and then hand over your banking information in order to “process the refund.” They might even ask you to hand over details about your PC or install additional software so they can access your system remotely. You see where this is going, right?

Just remember: No website can diagnose security problems on your PC remotely, and no legitimate app will tell you to just uninstall your antivirus software. 

Android Car Malware Spreads Through Built-In Updaters

Back when the “Internet of Things” was a buzzphrase, security experts warned that connecting virtually everything to the open web without protecting it would cause problems. After all, you don’t want your smart light bulbs to give away your Wi-Fi network password because it sends it in plain text, or your smart fridge to become part of a botnet that’s responsible for massive DDoS attacks, right? Well, if you have an Android Car device in your vehicle, that could be what’s happening. According to The Hacker News, back in June 2026, security researchers at Kaspersky uncovered a campaign targeting Android-based car head units with firmware developed by DoFun. The malware spreads through the same built-in software updaters designed to keep installed apps and the units themselves safe and secure. 

Once the device is infected, the malware downloads additional packages designed to join the device to a botnet used for malicious proxy services and ad fraud. Worse, because these devices often have their own internet connectivity for satellite navigation and emergency services, they can have persistent internet access, which makes them attractive and useful targets for hackers looking for devices they can use for their own purposes that also won’t draw much attention from their owners, and don’t benefit from protection from security software. 

Ask Our Expert: Is It Rude to Ask Someone to Take Off Their Smart Glasses?

Do you have a question about online privacy or security? I’m here to help! You can submit your question here, and I may answer it in an upcoming SecurityWatch column and newsletter. If you’re not subscribed to the newsletter, head here to sign up, and check back each week for the latest updates from PCMag’s security team. Now, on to this week’s question! 

Ruby L asks: Is it rude to ask someone to take off their meta glasses if we’re hanging out/talking to each other? That seems like an intrusion on the part of the wearer, but it also feels rude to ask.

Hi Ruby! Thanks for your question. I don’t think it’s rude at all to ask someone to take off their smart glasses at all when you’re hanging out! In fact, I’d reframe your question entirely: Instead of being worried that it’s rude to ask the wearer to take them off, I think it’s rude to keep them on, the same way we consider it rude if you’re out with a friend and they’re glued to their phone instead of actually talking to you.

If you trust the person not to record you while you’re hanging out, then maybe you can skip it. But if you don’t know them, or just want to be sure, I think it’s as fair as asking someone with a camera not to take your picture, or telling a friend who’s recording a group hangout that you’d rather not be in the video.

The responsibility should be on the wearer to disclose to those around them that they’re not recording or using them to photograph or film others inappropriately. In fact, Will Greenwald, our principal writer for consumer electronics, agrees with you. In his recent piece about how “pervert glasses” are ruining wearable tech, he notes that the real benefits of smart glasses have nothing to do with their cameras, and that he (and I, frankly) hope the category can take off without them.