Privacy in the AI Era Is Possible: What Proton’s CEO Says and How to Protect Yourself
AI tools are now part of daily life for millions of people. Whether you use ChatGPT, Google Bard, or any other large language model, you are feeding it personal information—emails you paste, documents you upload, questions about your health or finances. The convenience is real. So is the privacy risk.
Proton, the company behind encrypted email and VPN services, has been vocal about this tension. In a recent interview with Spiceworks, Proton’s CEO laid out what keeps him up at night when it comes to AI and privacy. His message is both sobering and pragmatic: privacy in the AI era is achievable, but only if consumers and companies take deliberate steps.
What Happened
Proton’s CEO explained that the biggest concern is not just that AI models might misuse your data—though that is a problem—but that the entire ecosystem is built on data collection habits we have normalized. Many AI services store your conversations, use them to train future models, and share them with third parties. The default setup for most free AI tools is that your input becomes part of the company’s training pool.
The interview highlighted one specific worry: the lack of transparency. When you use an AI tool, you rarely know exactly what data is collected, how long it is kept, or who else can access it. Terms of service often grant broad permissions. And because AI models can reconstruct patterns from training data, even anonymized inputs can sometimes be linked back to individuals.
Why It Matters
For everyday users, the implications are direct. If you paste a draft of a legal letter into a public AI chatbot, that text could become part of a training set and later appear in someone else’s output. If you ask an AI for personalized health advice, the details you share might end up stored on a server you have no control over. The convenience of AI can come at the cost of your digital privacy.
Beyond individual risks, there is a broader market shift. Proton’s CEO warned about vendor lock-in: once you rely on a particular AI provider for work or personal tasks, switching becomes hard because your data is tied to that platform. This mirrors what happened with cloud services, but AI lock-in can happen faster because models learn from your specific usage patterns.
The good news is that awareness is growing. More companies are offering privacy-first AI options, and regulations are starting to catch up. But waiting for laws to protect you may take years. In the meantime, you can act now.
What Readers Can Do
You do not need to stop using AI entirely. Instead, adopt a few practical habits:
Check the data policy before you hit send. Even free tools often have a setting to opt out of training. ChatGPT, for example, lets you disable chat history and model improvement in its settings. Do that before your first conversation.
Use incognito or temporary sessions where available. Some AI services offer a temporary chat that is not saved. Treat these like private browsing modes—helpful for one-off questions, but do not assume total anonymity.
Limit personal information. Treat AI conversations like public spaces. Do not share full names, addresses, phone numbers, account numbers, or other identifiable details. If you need personalized advice, use encrypted or local-first alternatives.
Choose privacy-respecting AI tools. Proton has its own AI product, Proton Scribe, which runs locally on your device or through end-to-end encrypted servers. Other options include Ollama for local LLM processing and open-source models like Llama that you can run yourself. Cloud-based tools that claim “zero-know encryption” are also worth considering—just verify their claims.
Encrypt your communications. For chats and emails that involve AI, use end-to-end encryption wherever possible. Proton Mail, for instance, encrypts messages automatically. Pair that with a good VPN (Proton’s is a solid choice) to hide your IP address from AI service providers.
Keep software updated. AI integrations are new; bugs and vulnerabilities will be found. Regular updates patch these holes.
Sources
- Spiceworks interview with Proton’s CEO: Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night (June 2026). Available via Google News.
- Additional context from Spiceworks articles on AI vendor lock-in and regulatory gaps (May–June 2026).
Protecting your privacy while using AI is not about paranoia—it is about being intentional. The tools are useful, but they should serve you, not the other way around. Take a few minutes today to review your AI settings and choose services that respect your data. That small effort can save you a larger headache later.