Safer Internet Day is a global event dedicated to promoting a more positive digital world, and for 2026, the campaign is focusing on one of the most pressing challenges in tech: “exploring the safe and responsible use of AI.”
Our interaction with the web has been transformed by the rapid rise of tools like ChatGPT, Gemini, and Grok. However, these advancements have arguably outpaced the industry’s sense of responsibility.
1. Lumo by Proton: The first privacy-first chatbot
Swiss-based Proton – the firm behind one of the best VPN and secure email services on the market – has long championed a private and free internet.
Since the launch of Proton Mai in 2014, the company has built a comprehensive security ecosystem designed to challenge Big Tech with privacy-first alternatives. In July 2025, Proton brought the same ethos to the generative AI space.
Lumo is Proton’s open-source AI assistant designed to match the functionality of other LLMs like ChatGPT without the privacy trade-offs.
Conversations are kept confidential through zero-access encryption, ensuring that no user data is used to train the underlying model or sold to third parties.
Speaking at the launch, Proton founder and CEO Andy Yen warned that Big Tech is using AI to “supercharge” the collection of personal data and accelerate the growth of “surveillance capitalism.”
Yen argued that AI should not become a surveillance tool, and said Lumo “puts people ahead of profits.”
Following Proton’s established model, Lumo offers both free and premium tiers. The Lumo Plus plan provides expanded features, including unlimited chats and support for larger file uploads.
The assistant is accessible via a web browser or through dedicated apps for Android and iOS.
2. ExpressAI: ExpressVPN’s upcoming private AI platform
ExpressVPN is also making a significant move into the sector with its new “private-by-design” platform, ExpressAI.
While originally scheduled for a February 5 launch, the provider has pushed back the release to refine the tool’s performance.
Despite the delay, we already have a clear picture of its architecture. Much like Lumo, ExpressAI uses zero-knowledge, end-to-end encryption, ensuring that conversations are unreadable to everyone.
It is also expected to include a disappearing messages feature, which allows users to set conversations to auto-delete after a set period. Furthermore, ExpressVPN claims that no user files are ever stored on its servers.
ExpressAI will be included with all ExpressVPN plans. Here are more details on everything it means for your VPN subscription.
3. From Meshnet to Threat Protection Pro: How NordVPN fights the danger of AI
Unlike Proton and ExpressVPN, NordVPN has yet to announce its own chatbot. Instead, the provider is focusing on helping users defend against some of the risks of the burgeoning AI agent ecosystem.
Much of this concern centers on OpenClaw, a viral open-source AI agent capable of executing code and accessing local files.
Recent research found that attackers uploaded at least 14 malicious skills to ClawHub, the public registry for OpenClaw extensions, with the goal of tricking victims into spreading malware.
NordVPN CTO Marijus Briedis says the most critical risk are “prompt injection” attacks. That’s because OpenClaw has access to your email and messages, allowing a malicious actor to craft content that hijacks the agent’s behavior. It’s essentially “weaponizing your own AI assistant against you,” Briedis said.
NordVPN’s Meshnet – a tool that lets you create a secure virtual network among up to 60 different devices at once – offers a straightforward solution to this.
Briedis said: “Meshnet can create a direct, encrypted connection between your devices without routing through third-party servers. You can access your self-hosted OpenClaw instance as if it were on your local network, while keeping it completely invisible to the public internet.”
The team at NordVPN has also been busy improving the functionality of its flagship Threat Protection Pro suite to defend users against AI-enabled scams.
In 2025, the provider launched tools that scan email links for threats in real time, block spam calls, monitor crypto scams, protect against session hijacking vulnerabilities, and even perform adult site blocking on mobiles.
One of the company’s top priorities for 2026 is to keep expanding the capabilities of the Threat Protection Pro beyond VPN protection, with a focus on building tools to help detect deepfakes and manipulated content.
Looking ahead
Safer Internet Day 2026 serves as a vital reminder that while innovation can fundamentally improve our lives, those benefits are easily undermined without robust safety and ethical frameworks.
While the broader AI industry has often struggled to keep pace with the risks inherent to its own platforms, several companies from the VPN sector are now stepping in.
From Proton and ExpressVPN’s private-by-design assistants to NordVPN’s proactive threat detection, these tools offer a blueprint for a more secure, AI-integrated future.
As these threats continue to evolve, the “arms race” between malicious automation and privacy-first defense will only intensify.
For users, the best defense is no longer just a strong password, but a comprehensive security suite built to navigate an automated world.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
https://cdn.mos.cms.futurecdn.net/2UMvPDp3snEwaGbRuCivjE-970-80.jpg
Source link
chiara.castro@futurenet.com (Chiara Castro)
