Privacy in the AI Era: One Key Worry from Proton’s CEO and How to Stay Safe

Generative AI tools like ChatGPT, Copilot, and Gemini are now part of daily life for millions. They help with writing, coding, research, and brainstorming. But as adoption surges, so do concerns about what happens to the data you feed into these tools.

Proton’s CEO believes privacy is still achievable in the AI era. But he also flags a specific risk that keeps him up at night. Here’s what that risk is, and how you can protect your personal data without giving up the convenience of AI.

What the Proton CEO Actually Said

In a recent interview with Spiceworks, Proton CEO Andy Yen acknowledged that privacy-preserving AI is possible, but pointed to a critical vulnerability: the lack of transparency around how major AI services collect, store, and use user inputs. The core worry isn’t just that companies might read your prompts — it’s that your conversations can be used to train models, included in data breaches, or shared with third parties without meaningful consent.

Yen’s real concern boils down to this: users have no reliable way to know what happens to their data once it’s sent to a cloud-based AI, especially when the provider offers no clear privacy guarantees or end-to-end encryption. And with AI models becoming more powerful, the volume of sensitive personal and professional data flowing into them is growing faster than safeguards can keep up.

Why This Matters for Everyday Users

If you use a free AI chatbot, there’s a good chance your prompts are stored, analyzed, and possibly fed back into training data — sometimes permanently. Even paid tiers may not guarantee that your data stays private. A single leak or a change in terms of service can expose months of conversations that include draft emails, medical questions, financial details, or proprietary work information.

What makes this different from typical cloud storage is scale. You might share far more details with an AI assistant than you would with a search engine or email provider, because the interaction feels conversational. The risk is both invisible and cumulative.

3 Practical Steps to Reduce Your Exposure

The good news is you don’t have to stop using AI. But you can choose tools and habits that give you more control.

  1. Use services that offer end-to-end encryption or on-device processing. Proton itself has launched an encrypted AI assistant (Proton AI) that processes requests without exposing your data to third parties. Other options include local-only models like Llama or Phi (run via apps like Ollama), or browser extensions that mask your inputs. If a provider can’t or won’t explain how your data is handled, that’s a red flag.

  2. Opt out of training data wherever possible. Many major AI services allow you to disable the use of your conversations for model improvement — but the setting is often buried or off by default. Look for privacy or data controls in your account settings, and turn off “Improve the model” or similar options. Some services (like ChatGPT) also let you delete your chat history. Do this regularly.

  3. Keep your personal and professional AI use separate. Don’t paste company documents, passwords, medical information, or sensitive client data into a free AI tool. Create dedicated accounts for work and personal use, and consider using a separate email alias or login to limit tracking.

3 Questions to Ask Before Using Any AI Tool

Before you sign up for a new AI service, run through this quick checklist:

  • Where is my data stored, and who has access to it? Look for a clear data processing agreement or privacy policy that states data isn’t sold or used for training without your consent.
  • Does the service support encryption in transit and at rest? End-to-end encryption means even the provider can’t read your messages. This is still rare in AI, but it’s the gold standard.
  • Can I delete my data permanently, or is it retained indefinitely? Check whether the tool allows you to view, export, and fully erase your conversation history. Some services keep metadata even after deletion.

The Bottom Line: Privacy Is Possible, But Not Automatic

Proton’s CEO is optimistic that privacy and AI can coexist, but his caution is worth taking seriously. The tools we use today are setting long-term precedent for data handling. By choosing services that respect your privacy and adjusting your own habits, you can enjoy the benefits of AI without handing over the keys to your digital life.

Sources:

  • Spiceworks, “Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night” (2026)
  • Proton official website and blog – product privacy features and data handling policies