Privacy in the AI Era: What Proton’s CEO Says Really Matters for You

Every time you type a question into ChatGPT, upload a document to an AI summarizer, or let an app rewrite your email, your data travels somewhere. The question is: where, and who else gets to see it? As more people adopt AI tools for work and daily life, the trade-off between convenience and privacy is becoming hard to ignore. In a recent interview with Spiceworks, Andy Yen, CEO of Proton (the company behind Proton Mail and Proton VPN), shared what keeps him up at night when it comes to AI and privacy—and offered some practical guidance for users who want both.

What Happened

In the interview, published June 4, 2026, Yen laid out his main concerns with the current AI landscape. He pointed out that many AI services are built on a business model that relies heavily on collecting user data—conversations, documents, even metadata about when and how you use the tool. That data is often stored in cloud servers that the company controls and can access, and in some cases it may be used to train future models without explicit user consent. Yen described this lack of transparency as one of the biggest risks in the AI era.

He also noted that the convenience of free AI tools often comes with hidden costs. Users might not realize that their prompts, uploaded files, or past conversations are retained and analyzed. For individuals in sensitive fields like journalism, healthcare, or law, that can be a serious problem. For everyday consumers, it means your private questions—about finances, health, relationships—could end up in a database you never agreed to.

Despite these concerns, Yen argued that privacy in the AI era is achievable. The key is choosing services that are designed with encryption and user control at their core, rather than as an afterthought. Proton itself is working on AI features that run locally or with strong end-to-end encryption, so users can benefit from AI without handing over their data.

Why It Matters

Yen’s message is important because the default assumption for many people is that “if it’s convenient, it must be safe.” That’s rarely true with free online services. In 2026, AI tools are embedded in everything from search engines to writing assistants to photo editors. Most of them are not private out of the box. Even when companies claim to anonymize data, the methods can be incomplete, or the privacy policy may allow sharing with third parties.

The risk isn’t just theoretical. We’ve seen cases of AI chatbots leaking conversation histories, cloud storage being breached, and user data being used to train models that later reproduce personal details. For consumers, the consequences can range from embarrassing leaks to identity theft.

Proton’s approach—encrypting data so that the company itself cannot read it—offers a different model. But not everyone will switch to Proton. The broader lesson is that users need to become more aware of what happens to their data when they use an AI tool. Just because a tool is popular doesn’t mean it respects your privacy.

What Readers Can Do

You don’t need to be a cybersecurity expert to reduce your privacy risk with AI. Here are concrete steps you can take starting today:

  • Audit the AI tools you use. Check their privacy policies. Do they store your conversations? Do they use your data to train models? If the policy is vague or says they “may” share data, treat it as a red flag.

  • Prefer services with end-to-end encryption. For email, cloud storage, and messaging, choose providers that encrypt data before it leaves your device. Proton is one example, but there are others like Tutanota or Cryptee. For AI chat, look for tools that run locally on your device or that clearly state they don’t log your inputs.

  • Avoid sharing sensitive personal information with AI chatbots. Even if you trust the service, it’s wise to treat any chat AI like a public conversation. Don’t paste your passwords, Social Security numbers, or confidential documents unless you are absolutely sure about the data handling.

  • Turn off chat history and training features. Many AI services, like ChatGPT and Google Gemini, allow you to disable conversation storage. This reduces the chance that your data will be used for model training. It’s usually found in the settings or privacy menu.

  • Use a VPN when accessing AI tools from public Wi-Fi. This doesn’t prevent the AI company from seeing your data, but it hides your IP address and location from your internet provider and anyone on the same network.

  • Consider running open-source AI models locally. If you have a decent computer, you can run models like Llama or Mistral on your own machine. No data ever leaves your hard drive. Tools like Ollama make this easier.

Sources

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