Privacy in the Age of AI: Proton’s CEO Explains the Real Risk

The rapid adoption of AI tools over the past year has reshaped how people work, search, and create. But for many users, the excitement is tempered by a growing unease about what happens to the data they feed into these systems. In a recent interview with Spiceworks, Proton’s CEO offered a blunt assessment of where we stand—and what genuinely worries him.

What Happened

Proton, the company behind Proton Mail, Proton VPN, and other privacy-first services, has been vocal about building secure alternatives to big tech. In the interview (published June 4, 2026), the CEO stated that privacy in the AI era is achievable if products are designed with encryption and user control from the ground up. However, he pointed to one overriding concern: the sheer scale of data being collected and processed by today’s AI models, often without meaningful consent or oversight.

The specific thing that keeps him up at night, according to the article, is the risk that AI systems could become a central point of failure for personal privacy. As more companies rush to embed AI into their products, the potential for sensitive data to be used for unintended purposes—such as training, profiling, or sharing with third parties—grows exponentially.

Why It Matters

Most consumers don’t realize how much of their input to chatbots, image generators, or writing assistants is being stored, analysed, and sometimes reused. Free AI tools often rely on user data to improve their models, a trade-off that is rarely transparent. The CEO’s concern reflects a broader pattern: we are trading convenience for control, often without a clear picture of the long-term consequences.

For the average person, this isn’t just an abstract debate. If you’ve ever pasted a draft email, a health question, or a financial document into an AI tool, that data may have been retained or used to train the next version of the model. Even when companies promise not to store inputs, enforcement is inconsistent and hard to verify.

Proton’s approach—end-to-end encryption for AI interactions—is one possible solution, but it remains niche. The CEO argues that privacy should not be a premium feature; it should be the default.

What Readers Can Do

You don’t have to stop using AI altogether, but you can take practical steps to reduce your exposure.

  • Check data policies. Before using any AI tool, read how your inputs are handled. Look for services that explicitly state they do not store or train on your data. Avoid platforms that treat your text as their training material unless you are comfortable with that.
  • Limit what you share. Assume that anything you type into a public AI interface could be seen by someone else. Avoid sharing personal identifiers, medical details, financial information, or anything you wouldn’t want posted online.
  • Use privacy-focused alternatives. Services like Proton’s own offerings (when available) or local open-source models give you more control. Running a smaller model on your own device eliminates the need to send data to a server.
  • Encrypt where you can. For communication and storage, end-to-end encryption remains the gold standard. It doesn’t protect you from what you type into an AI, but it limits the reach of any breach.
  • Keep software updated. AI platforms are still evolving, and security patches matter. Enable updates to protect against vulnerabilities that could leak your data.

None of these steps is foolproof, but together they significantly reduce the surface area for privacy erosion.

Sources

  • “Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night.” Spiceworks, June 4, 2026.
  • General knowledge about AI data practices from public reports and consumer advocacy groups.