This week, PCMag’s security team traveled to Las Vegas for the Black Hat cybersecurity conference to engage in some intense cybersecurity conversations about adversarial AI, malware, and all the ways to stop bad guys from ruining your business or scamming you out of house and home. We saw everything from potentially explosive EV chargers to global malware campaigns run by the mob. Here’s a summary of some of the most interesting briefings, demonstrations, and panels we attended during the show.
Did You Click on an Ad Owned by Russian Mobsters?
A particularly interesting briefing came from the team at InfoBlox. Researchers at the threat intelligence firm presented evidence that an organized crime group known as VexTrio is operating a traffic distribution system (TDS) that intercepts consumers’ browsing traffic, and serves up fake alerts, downloads for fake dating apps, and prompts people to download malware disguised as antivirus software.
Dr. Renee Burton is part of the Infoblox research team, and she recommends avoiding these scams by simply not allowing notifications for websites while browsing. She also suggests researching apps before downloading and ignoring device infection pop-ups or alerts that you may see while browsing.
Learning How to Take AI Down a Peg or Two
In dystopian fiction, humanity gives more and more control to AI and then suffers the consequences. As an Israeli research team reported, that’s not fiction anymore. In a presentation frequently interrupted by applause, they demonstrated simple techniques for subverting Google’s Gemini through calendar invitations and even through email.
The researchers forced Gemini to delete messages, spew invective at the user, steal email messages, and even use Google Home to open windows. Google has added defenses against these targeted “promptware” attacks, but the team warns that it anticipates promptware attacks to increase in complexity and severity.
In a separate presentation on Thursday, researchers from the University of Central Florida and the Cognitive Security Institute announced their predictions for the next two years of global AI adoption. Those predictions included increasingly sophisticated catfishing scams, more effective phishing attempts, and, concerningly, everyone will have a so-called digital twin created using their online content.
Researchers at Black Hat say that AI is here to stay. (Credit: Kim Key)
To back up these predictions, the researchers showed a version of a “digital twin” of Chief Justice John Roberts created by their students. The large language model appeared to have been trained on Roberts’ public Supreme Court opinions and rulings, making it a not-so-fun and not-so-convincing fake, but their point was still clear: AI is here to stay, so it’s wise to learn how to live with it. Below is an excerpt from a conversation we had with the bot version of Chief Justice Roberts.
A conversation between a PCMag analyst and an LLM trained to act like a Supreme Court judge. (Credit: Habit/PCMag)
AI is trained to mimic and satisfy the humans it interacts with, so don’t be surprised if someone uses it to trick your family and friends. The team suggested designating a “safe word” with your family members and friends, so if, or when, your digital twin comes along, your loved ones won’t become its victims.
Using AI to Fight Malware
Yes, AI produces deepfake videos and convincing scams, but it can also help fight those threats. For example, every day, thousands of people fall for one trick or another and infect themselves with infostealer malware. The malware sucks up bitcoin wallets, password stashes, and all sorts of personal data, then sends an info package back to its maker.
Flare.io researcher Estelle Ruellan gained access to millions of these info packages and devised a system using two very different AIs to analyze them and return actionable details about the malware campaigns that distribute these attacks. This project specifically focused on screenshots included in the packets, but Ruellan and the Flare.io team plan a deeper analysis using the entirety of the captured info packages.
Get Our Best Stories!
Stay Safe With the Latest Security News and Updates
By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy Policy.
Thanks for signing up!
Your subscription has been confirmed. Keep an eye on your inbox!
Examining An Ever-Evolving Cyber Threat Landscape
Speaking of malware, one of cybersecurity’s pioneers provided a malware history lesson as part of the opening keynote for the event. Mikko Hypönnen, a longtime cybersecurity researcher who worked with F-Secure and WithSecure, opened his speech by recounting his days of keeping viruses and worms, which were made by random kids as pranks, on floppy disks. He told us that money is the root of all evil, permeating the cybercrime landscape. Once criminals, adversarial governments, and everyone else learned that they could use malware to steal money, the stakes got a lot higher for cybersecurity professionals.
Mikko Hyponnen onstage for the first keynote at Black Hat 2025 (Credit: Kim Key)
Hypönnen said that AI can assist in the fight against today’s malware, and companies and governments should look into using AI-enhanced tools to protect people. There are a lot of software vendors at Black Hat, so Hypönnen directed comments at them in particular, urging companies to stop putting the onus on consumers to protect themselves from scams. Instead, they should try making software and systems that are more inherently secure.
Hypönnen ended the chat by announcing his departure at the end of the week from the cybersecurity industry. He’s getting into anti-drone tech, which is quite a change!
Recommended by Our Editors
Hacks, Hacks, and More Hacks
Any hacking group that wants to penetrate your network typically starts small, sneaking in an app that can then expand its control. Then they need remote communication with that app, and its commands can reveal the attack to network security.
Adam Crosser of Praetorian found a way to effectively make Zoom or Teams do the work of evading security, and to disguise those remote control messages as web conferencing traffic. Like any ethical hacker, he notified Zoom and Microsoft. Zoom has locked down its tech to prevent this attack, but Microsoft hasn’t.
In a different demonstration, researcher David Moshe and his team discovered that security cameras can leave networks wide open to hackers. Moshe and Claroty’s Team82 found a service on the web called axis.remoting and parlayed that discovery into an attack that gave them full control of any security cameras and camera control servers from Axis Communications. Axis fixed the problem quickly, but teams like Moshe’s (and their equivalents on the Dark Side) continue to seek and find security holes.
Panelists gather for a group photo after a presentation (Credit: Kim Key)
Is Cybersecurity Your Next Big Career Move?
Finally, a panel of successful women, each of whom worked in the industry for a least a decade, offered some solid suggestions for navigating the industry as visible minorities. The group answered questions from the crowd about dealing with uncomfortable office politics and staving off the dreaded imposter syndrome.
The panelists told us that some of the keys to entering the industry and staying employed are as follows: keep your skills sharp, stay curious, seek out a community of like-minded and empathetic mentors or fellow professionals, be yourself, and stay flexible when it comes to taking jobs that are off the beaten path.
About Neil J. Rubenking
Principal Writer, Security
Read the latest from Neil J. Rubenking
About Kim Key
Senior Writer, Security
