ZDNET’s key takeaways
- NordVPN develops its products for AI threats.
- An invisible, catch-all solution is the easiest to adopt.
- The future of secure AI is small, local models.
AI has hijacked many fundamental principles of the security landscape. The cyber security industry is adapting to this threat by preparing for new threats and implementing AI as a tool.
Redefining Antivirus
NordVPN recently launched what it calls “Next-Gen Antivirus,” or NGAV, as a catch-all security tool for your online life. The offer tries to build on antivirus, a now incomplete term that refers to a time when threats came in the form of malicious files. But files are no longer the main problem. Socially engineered scams, phishing, and identity theft have ballooned, becoming a significantly larger part of the threat landscape and rapidly scalable with AI.
Also: Why I still recommend NordVPN for most people in 2026
“We don’t believe that viruses now come through floppy disks and CDs,” Briedis said. “All the protection you need right now, basically, is on the Internet.”
Next-Gen Antivirus is NordVPN’s answer to this reality. Getting started with its VPN offering and securing the “tunnel” was the company’s first task. The second, as Briedis said, was about what comes through that tunnel. NGAV, which NordVPN describes as “a security system for your entire digital life,” promises to warn users about scam websites, block ads and trackers, and remove malware “quietly in the background.”
The last part is crucial. As AI-enabled threats multiply and become more seamless, Briedis believes that a solution that runs invisibly in the background and requires no technical expertise to manage is the best way to convince people to adopt better security hygiene.
Briedis echoed what Proton CEO Andy Yen told ZDNET in April: Security companies can offer users protection, but they can’t save them from themselves. In Briedis’ case, the feeling shows how easy – and risky – trying out AI tools has become.
Also: Rogue AI incidents hit ‘tens of thousands’: Can businesses trust these tools?
“There’s always a human element,” he said. “We usually joke in the IT world, among the geeks, that the weakest link between the computer and the chair is the human.”
Briedis said NordVPN still thinks it’s important to help people understand the tools they use, where encryption matters, the data they share and with whom, and where the risks lie.
Agents as the next attack surface
But with the rise of privacy-infringing personal AI agents, the Internet isn’t really the only place where users need protection. How does NordVPN square the reality of agents operating on devices, with or without the Internet, as an area it cannot control? Rather than stopping at computers and humans, Briedis said the key is making agents themselves less vulnerable.
“We are looking at how to protect agents as well, because we see that agents [themselves] could be attacked on your computer,” Briedis said. “I think that’s where all cybersecurity ultimately goes. It’s kind of a no-brainer, because we’re going to use it [agents] every day.”
Also: Inside Google’s faster Chrome patch strategy to block AI attacks on your browser
This feature is still in development and doesn’t have a name yet, but it would be part of the NordVPN app when it’s released.
Use AI in NordVPN
Like most tech companies, NordVPN has incorporated specialized AI into its products, while also using internal frontier models from large labs.
For its next-generation antivirus feature, NordVPN said it is developing small models internally to tackle specific tasks, such as spotting phishing attempts and other scams.
“There are about 20 parameters that we can tweak on the fly, and they’re working pretty well right now,” Briedis said. These models can also run locally on a user’s device. “You don’t need a lot of hardware to do that these days because they’re really small.”
Also: A low-tech solution from the past can be your best defense against AI deepfakes
As for the use of AI internally, Briedis said that NordVPN builds with a mix of models, but now often favors open-weight models. Like many companies, it started out mainly using frontier models from OpenAI and Anthropic, but costs quickly built up. With where the open-source model performance is now, NordVPN doesn’t lose any capabilities by switching.
As a result, Briedis said the amount of code produced at NordVPN “has increased.” However, that code goes through a thorough review process that can create an approval fee.
“We don’t do the simple software development life cycle. We make sure the software development life cycle,” he said.
Still, there are limitations: Briedis said one of NordVPN’s “non-negotiables” is that teams handling payments, for example, don’t use AI at all. Experimentation and innovation are reserved for other parts of the business.
Also: I tested NordVPN’s free scam checker with real phishing emails
Briedis also noted that NordVPN programmers experience a now-common reality across tech companies: the pressure to adapt as their careers shape changes.
“I think every software development company is currently changing how they work, and many programmers [are] trying to figure out what they’re going to be doing now in five to 10 years, because their profession is changing so much,” he said.
Privacy and AI on Chance
Given the growing unpopularity of AI and its proven privacy risks, some security-conscious customers may feel uneasy about AI and privacy tools sharing space in a product. But Briedis said NordVPN has yet to experience customer pushback over AI in features like next-gen antivirus, in part because AI use is already common.
“I can’t imagine people not using AI at all these days,” he said.
Also: 7 privacy settings I change ASAP on every new phone – here’s why
Furthermore, the main way to maintain trust with your customers is not to use their data for model training, which Briedis reiterates that NordVPN stands for.
Protections like the EU’s GDPR can add a layer of insurance on top of private company promises. On that front, however, Briedis believes that over-regulation – which he believes has happened in Europe – risks stifling innovation.
“For example, the CRA law that we will have in Europe – it will be really difficult for us to create software if we want to tick all the boxes to develop through this act,” he said.
The future of safe AI use
ZDNET asked Briedis about his dream security tool for the AI-embedded future. What would he build if he had no limits and could seamlessly address any attack surface?
“For now, I think it’s all about the information that’s happening around you in your environment,” he said. “It could be an app or device that tracks you and tracks all your information, with your name and what’s happening on the internet around you as a person, to make sure the data is linked together, safe and secure.”
It makes sense: something to monitor all your exposure points, but for your protection rather than exploitation. Ideally, a product like this would protect your information and automatically delete it from the Internet as needed.
Also: 10+ VPN tricks and tips I recommend to everyone
“The ideal scenario, I would say, is everything on premises and on your machine, but that’s not going to happen,” Briedis added. “But I think this tool should help you think about making sure your data is encrypted at rest—even maybe encrypted in use, which is a really hard task if you don’t have specialized hardware these days—and also encrypted in transit. The three pillars overall, I think, would be the best-case scenario.”
Consumer hardware such as phones and PCs are adapting to accommodate on-device AI features. This capability does not mean that Briedis sees NordVPN and the hardware desperate to manage the attack surface anytime soon.
“It’s never been our strategy to go into hardware, and I don’t think we’ll do it in the next couple of years, that’s for sure,” he said. “But you never know, you know? Maybe there’s a new, really affordable chip that you can use to create a device.”
For a more immediate outlook on the future of AI and security, however, Briedis thinks the key is still getting smaller and staying local.
“I think in the end, tech-savvy people go to their own LLMs, on their own computers,” he said. “It’s just a matter of time in my mind.”