Can You Keep Your Privacy in the AI Era? Proton’s CEO Shares His Biggest Concern

Can You Keep Your Privacy in the AI Era? Proton’s CEO Shares His Biggest Concern

Introduction

Using AI tools like ChatGPT, Gemini, or Copilot often feels like a trade‑off: you get fast answers and smart suggestions, but you hand over your questions, documents, and sometimes personal data. For privacy‑conscious users, that exchange feels increasingly uneasy. As more people rely on generative AI for work, research, and everyday tasks, the question isn’t just whether AI can be useful—it’s whether you can use it without losing control over your information.

What Happened

In a recent interview with Spiceworks (June 2026), Proton’s CEO discussed the state of privacy in the age of AI. Proton is best known for its encrypted email and VPN services, so its CEO’s perspective carries weight in the digital privacy community. The conversation covered both the promise and the peril of current AI tools. According to the interview, the CEO believes privacy is still achievable as AI becomes more embedded in daily life, but he pointed to one specific threat that worries him most: the casual, often invisible use of user conversations to train and improve AI models.

The CEO argued that while several AI providers have improved their transparency in recent months, the default behavior of many popular tools remains data‑hungry. Without deliberate user action, prompts and uploaded files can be stored indefinitely and used for model training, even if the user later deletes the chat.

Why It Matters

This is not a theoretical risk. If you use a free or basic tier of a major AI chatbot, your conversation data is likely being used to refine the system. In some cases, that data may include personally identifiable information or sensitive business details. Even when companies promise not to use data for training, the lack of end‑to‑end encryption means the provider could still access your chats if compelled by law or in the event of a breach.

The implications go beyond corporate policies. For many users, the convenience of AI leads to oversharing. People paste private documents, ask about health issues, or discuss financial plans inside a chat window, assuming it’s as private as a search engine. It’s not. The CEO’s concern reflects a broader pattern: as AI tools become more capable, the line between “public” and “private” conversation blurs, and most users aren’t aware of how their data is handled.

What Readers Can Do

There are practical steps you can take right now to reduce your exposure without abandoning AI tools.

1. Read the privacy policy (or at least the relevant parts).
Most AI services have a section that explains whether your conversations are used for training. For example, OpenAI allows users in some regions to opt out of training, but the default is often opt‑in. Check the settings page of any tool you use regularly.

2. Use privacy‑focused alternatives.
Consider AI tools that were built with privacy as a core feature. DuckDuckGo’s AI chat and Brave’s AI assistant, for instance, do not store conversations. Proton itself offers an encrypted AI service called Proton Scribe (integrated with Proton Mail), which keeps data on‑device or in end‑to‑end encrypted storage. Local models like Llama 3 or Mistral can be run entirely on your computer without sending data anywhere.

3. Never paste sensitive information into a generic chatbot.
If you wouldn’t put it in a public forum, don’t put it in a chat with a free AI tool. That includes full names, addresses, account numbers, health records, or trade secrets. Use dummy data or redact details when testing ideas.

4. Clear your conversation history regularly.
Most platforms allow you to delete chats. Make it a habit to clear them after you’re done, especially if the service stores them on its servers. Some tools also let you disable chat history altogether—enable that option if available.

5. Be especially careful with workplace AI.
If your employer provides you with a corporate account for an AI tool, assume every prompt is visible to the company. Treat it like work email. Do not use it for personal matters unless your employer explicitly says it’s private.

Sources

• Spiceworks interview with Proton’s CEO, June 2026. Original article: “Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night.” (Note: direct quotes from the interview were not independently verified for this piece.)
• DuckDuckGo AI Chat privacy policy: duckduckgo.com/privacy
• Brave AI privacy information: brave.com/leo
• Proton’s approach to AI: proton.me