Privacy in the AI Era Is Possible – But Here’s What Worries Proton’s CEO
If you’ve used ChatGPT, Google Gemini, or any other large language model in the past year, you’ve probably asked yourself: Is my data safe? It’s a fair question, and one that Andy Yen – the CEO of Proton, the company behind Proton Mail and Proton VPN – addressed in a recent interview with Spiceworks.
His message was measured: privacy in the age of AI is achievable, but it’s not automatic. And there’s one thing, in particular, that keeps him up at night.
What happened
In an article published June 4, 2026, Yen acknowledged that the tools exist to protect your data while using AI services. Proton, after all, has built its reputation on encrypted email and a zero-access architecture. But he also pointed to a persistent risk that even the best technology cannot fully solve on its own.
The concern isn’t about encryption breaking or some novel vulnerability. It’s simpler and harder to fix: many people are handing over highly sensitive information to AI systems without understanding how that data is used, stored, or shared. The CEO’s specific worry is that users will treat AI chatbots as private by default, when in reality most mainstream AI tools log conversations and may use them for model training unless you opt out – and sometimes even then.
Why it matters
AI is now embedded in everything from writing assistants to search engines to customer support. The convenience is real, and the adoption is accelerating. But the privacy implications are not always obvious. When you paste a draft of a legal contract, a medical history, or a confidential business plan into a chatbot, that information leaves your control.
Even if a company promises not to misuse your data, the risk of a breach or a subpoena remains. And not every provider is transparent about what they do with your inputs. Some have changed their privacy policies after launch, and others have been caught using user data for training without explicit consent.
For everyday users, this creates a gap between expectation and reality. We assume a private conversation with an AI is just that – private. But in many cases, it’s closer to shouting across a public square.
What readers can do
The good news is that you don’t have to stop using AI tools. You just need to be more deliberate about how you use them. Here are a few practical steps based on the kind of advice Proton’s CEO and other privacy advocates often recommend.
Use privacy-focused AI services. Several companies now offer AI chat tools that are built with encryption and no-logging policies. Proton itself is developing its own AI writing assistant with a zero-access architecture – meaning even the company cannot see your data. Other independent services exist that process queries locally on your device or allow you to control data retention.
Avoid sharing personal or sensitive information. Even with a trustworthy provider, it’s wise to treat any AI conversation the same way you would treat a post on a semi-public forum. Don’t include passwords, Social Security numbers, medical details, or trade secrets. If you need to analyze a document, use a tool that runs entirely on your own machine.
Check privacy settings and policies. Before you start using a new AI service, spend five minutes reading its data handling section. Look for answers to questions like: Are my conversations stored? Do they use my data to improve the model? Can I delete my history? If the answers are unclear, consider that a red flag.
Consider local models. For users who are especially privacy-conscious, running an open-source AI model on your own computer is increasingly practical. Tools like Llama, Mistral, and others can be downloaded and run offline. You lose the convenience of a cloud-based service, but you gain full control over your data.
Enable encryption where available. Some AI services offer end-to-end encryption for conversations. It’s not common yet, but it’s becoming more available. Use it when you can, and remember that even with encryption, you still need to trust the client software on your device.
Sources
The insights from Proton’s CEO discussed here come from an interview published by Spiceworks on June 4, 2026. The article is titled “Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night.” Proton is known for its encrypted email service, VPN, and privacy-focused cloud storage.
No single tool or habit will make your data completely invulnerable, but a combination of awareness and the right technology goes a long way. As Yen put it, privacy in the AI era is possible – but it requires you to stay awake to what your AI tools are actually doing with your data.