How to Protect Your Privacy in the AI Era: Tips from Proton’s CEO
In a recent interview with Spiceworks, Proton CEO Andy Yen made a straightforward claim: protecting your privacy in the age of artificial intelligence is still possible. But he also admitted that one specific AI risk keeps him up at night. That risk, he explained, is not the fear that AI companies will sell your data, but something more insidious: the ability of AI to generate convincing deepfakes using personal information people already share freely online.
Yen’s comments are worth paying attention to. Proton has built a business on encrypted email, VPN, and cloud storage services that put user privacy first. If anyone has a realistic view of what it takes to stay private in 2026, it’s him.
What happened
The interview, published on June 4, 2026, covered a range of privacy topics, but the part that stood out was Yen’s biggest worry. He pointed out that AI tools are now powerful enough to turn a few seconds of someone’s voice or a handful of publicly available photos into a realistic fake video. And the risk is not only for celebrities or politicians, but for ordinary people. Once a deepfake is out there, it can be used to impersonate you, damage your reputation, or trick your family and colleagues into sending money or sharing sensitive information.
The problem is compounded by the fact that many AI applications are built on top of massive datasets that users contribute to unknowingly. Every prompt you type into a public chatbot, every image you generate, every conversation you have with a voice assistant – all of that data may be stored, analyzed, or reused in ways you never agreed to.
Why it matters
You might think you have nothing to hide, but that misses the point. The real issue is consent. Your voice, your face, your writing style – these are personal identifiers that you never explicitly handed over to anyone. Yet they are being scraped from social media, forums, and even private messages, then fed into AI models that can mimic them with startling accuracy.
For consumers, the practical consequences are immediate:
- A deepfake voice call from a “relative” asking for an emergency payment.
- A fake video of you saying something damaging, used for extortion or harassment.
- An AI impersonating you in a customer support chat to reset passwords.
These are not hypothetical scenarios. Reports of voice-cloning scams have been rising for years, and the tools to create them are cheaper and more accessible than ever.
What readers can do
The good news is that you don’t have to stop using AI to stay safe. Here are concrete steps you can take, based on the advice that Yen and other privacy experts have offered:
1. Use encrypted communication services. When you send an email or a file, make sure it is end-to-end encrypted. Proton Mail, for instance, encrypts your messages so that even the provider cannot read them. The same principle applies to messaging apps like Signal. This prevents your data from being intercepted or scraped by third-party AI models.
2. Limit what you share with AI chatbots. If you use a chatbot like ChatGPT, Gemini, or Copilot, assume that everything you type could be stored and used for training. Do not paste in sensitive documents, passwords, or personal stories that you would not want to become public. If you need to use AI for private tasks, look for tools that run locally on your device or that have clear privacy policies promising not to retain your inputs.
3. Be careful with voice and video. Think twice before recording and sharing audio clips of yourself, even in private group chats. A few seconds of your voice is enough to create a credible clone. Similarly, limit the number of high-quality photos and videos you post publicly. The less raw material available, the harder it is to impersonate you.
4. Use a VPN to mask your IP address. A VPN, like Proton VPN, hides your real location and makes it harder for AI services (and the trackers they embed) to build a profile on you. Combined with a privacy-focused browser, this reduces your data footprint significantly.
5. Be on the lookout for deepfakes. If you receive an unusual request from a friend or colleague, especially involving money, verify it through a separate channel. Call them back on a phone number you already know. Ask a question only they would answer. Do not trust a video or voice clip alone.
6. Use strong, unique passwords and enable two-factor authentication. This one is timeless, but especially important now. If an AI manages to impersonate you, a good password and MFA can stop the attacker from actually accessing your accounts.
The bigger picture
Yen’s main point is that privacy in the AI era is not impossible – but it does require deliberate choices. You cannot rely on companies to protect you by default. The market incentives are still tilted toward collecting as much data as possible.
Proton itself is working on AI tools that run entirely on the user’s device, so that no data ever leaves your control. Whether those products will arrive in time to match the convenience of cloud-based AI remains to be seen. But the approach aligns with what Yen calls a “privacy-first” future, where the user – not the service provider – holds the keys.
For now, the most practical thing you can do is to treat every interaction with an AI tool as a public act. Assume the data you share may live forever, and plan accordingly.
Sources
- Spiceworks interview with Andy Yen, June 4, 2026: “Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night”