Privacy in the AI Era Is Possible—But One Risk Worries Proton’s CEO Most

AI tools like ChatGPT, Google Assistant, and smart note-taking apps are becoming as common as email. They promise convenience, but they also require trust—trust that your conversations, documents, and habits aren’t being collected, analyzed, or sold.

Proton’s CEO Andy Yen recently told Spiceworks that privacy in the age of AI is achievable, but he also pointed to a specific threat that keeps him up at night. According to the interview, Yen’s concern isn’t about AI itself, but about the unchecked data aggregation that many AI services rely on—and the growing risk that this data could be accessed by governments or malicious actors. While the full interview wasn’t available for review, the pattern is familiar: the more data you feed into AI tools, the bigger the target you become.

This isn’t an abstract problem. Every time you paste a sensitive email into an AI chatbot or let a voice assistant listen to your conversations, you’re handing over information that could be stored, analyzed, or leaked. Yen’s warning is that without strong encryption and clear data policies, the very tools designed to help you could become a privacy liability.

Why It Matters

The convenience of AI often comes with hidden costs. Many free AI services are funded by data collection—your prompts, your uploads, even your usage patterns become part of a training set or a profile. If a company’s servers are breached, or if a government subpoenas your chat history, you may have little recourse.

Proton’s own services (email, VPN, drive) use end-to-end encryption, meaning even Proton can’t read your data. This is the standard more consumers should demand from AI tools, but few currently meet it. The CEO’s worry is that most people don’t realize how much data they’re giving away until it’s too late.

What You Can Do Today

Privacy in the AI era doesn’t require you to stop using these tools. But it does require being intentional. Here are a few concrete steps:

1. Use privacy-first AI alternatives.
Consider tools that process data locally or use end-to-end encryption. For example, running a local large language model (like Llama or Mistral) on your own machine keeps your data offline. Open-source alternatives like PrivateGPT or Ollama let you chat without sending anything to a cloud server. For tasks that need cloud AI, choose providers with clear no-log policies and independent audits.

2. Turn off data sharing settings.
Most AI assistants—whether from Google, Apple, or Amazon—default to sharing data to improve the service. Go into your settings and disable options like “improve by sharing recordings” or “allow human review.” This can dramatically reduce what’s stored about you.

3. Avoid pasting sensitive information into chatbots.
It’s tempting to paste a draft of a contract or a private email into ChatGPT for feedback. Instead, anonymize the text first. Remove names, addresses, and any identifiers. Even better, use a tool that lets you run the AI locally.

4. Review app permissions regularly.
Many AI-powered apps request access to your microphone, camera, and contacts. Take five minutes each month to review which apps have permissions and revoke those that don’t need them. On iOS and Android, you can set permissions to “ask every time” for sensitive data.

5. Enable end-to-end encryption wherever possible.
If you’re using a service that offers it—like Proton’s suite, or Signal for messaging—turn it on. For AI tools, ask whether the service encrypts your data both in transit and at rest. If the answer isn’t clear, consider that a red flag.

The Bottom Line

Yen’s message is that privacy and AI aren’t opposites. The threat he worries about—unchecked access to user data—is real, but it’s also manageable. It starts with choosing tools that respect your privacy and adopting small habits that limit how much you expose.

You don’t have to abandon AI to keep your data safe. You just need to know where the risks are and how to sidestep them.

Sources:

  • Spiceworks, “Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night,” June 4, 2026.