Protect Your Privacy in the AI Era: What Proton’s CEO Wants You to Know
Proton CEO Andy Yen recently told Spiceworks that privacy in the age of AI is not only possible but necessary—yet he admits one thing keeps him up at night. It’s not the technology itself, but how most companies are building it: by harvesting vast amounts of user data to train models, often without meaningful consent. For anyone who uses AI tools like ChatGPT, Google Bard, or even voice assistants, the question isn’t whether AI is useful. It’s whether you can use it without giving up control of your personal information.
The answer, according to Yen, is yes—if you make deliberate choices about the services you rely on and how you use them.
What happened: the CEO’s core concern
In the interview, Yen pointed out that the current AI boom runs on a data-hungry model. Free AI tools are rarely free: they collect what you type, analyze your behavior, and often feed that data back into training pipelines. Even with promised privacy controls, users have limited visibility into what companies actually do with their inputs. Proton’s CEO described this as a “surveillance-driven AI” approach—one that mirrors the ad-targeting models we’ve seen from big tech for years.
The one thing that keeps him up at night is the possibility that this model becomes the default, normalizing the idea that privacy is a trade-off for AI convenience.
Why it matters for everyday users
Most consumer AI tools today are built by companies whose primary business is data monetization. When you ask a chatbot a personal question, upload a document for summarization, or use an AI writing assistant, you may be handing over sensitive information—emails, financial details, health notes, or confidential work files. These data points can be stored, analyzed, and used to improve the very system that later serves ads or influences recommendations.
And once your data enters a model, it’s difficult to remove it. Even companies that allow you to delete your account rarely guarantee that historical training data is purged.
For privacy-conscious individuals, the real risk isn’t a single breach but the cumulative erosion of control: everything you share becomes a permanent input into a system you don’t fully understand or govern.
What readers can do: actionable steps
You don’t need to stop using AI. You just need to change how you approach it. Here are practical steps, based partly on the approach Proton advocates, and partly on general best practices.
1. Use encrypted communication services
Whenever you share sensitive information—via email, file transfer, or messaging—ensure it’s protected by end-to-end encryption. Proton’s own services (ProtonMail, ProtonVPN, ProtonDrive) are built with zero-access architecture—meaning even Proton cannot read your data. Other trustworthy encrypted options include Signal for messaging and Tresorit for file storage.
2. Avoid free AI tools that monetize data
If a tool is free and doesn’t come from a company with a clear privacy-first business model (like a subscription), assume your data is the product. Consider paid versions that offer privacy guarantees, such as ChatGPT Plus with no training on your inputs, or privacy-focused alternatives like Brave’s AI assistant or DuckDuckGo’s AI chat.
3. Isolate AI usage from your real identity
Create a separate email account (preferably encrypted) for signing up to AI services. Use disposable or alias emails when you need to test a tool. Turn off conversation history or opt out of model training in the settings of tools you do use. Most major providers offer these controls—though they may be buried in menus.
4. Limit what you share
Treat AI tools like a public forum: don’t paste personal data, passwords, or confidential documents unless absolutely necessary. If you need AI help with a work document, strip out identifying details first. Think of your inputs as something that could be stored indefinitely.
5. Use a VPN to obscure your IP and location
Many AI services track usage sessions by IP address and device fingerprints. A good VPN (like ProtonVPN, which has a free tier) hides your real IP and encrypts your connection, making it harder to link your AI activity to your identity.
6. Read privacy policies—at least the summary
Before using any AI tool, check how the provider handles your data. Look for statements on data retention, third-party sharing, and whether your inputs are used for model training. If the language is vague or if the company has a history of data misuse, think twice.
Sources
The remarks from Proton CEO Andy Yen were originally reported in an interview with Spiceworks:
Privacy in the AI era is possible, says Proton’s CEO, but one thing keeps him up at night.
Privacy in the AI era isn’t a lost cause. It just requires a shift from passive consumption to intentional tool choice. The CEO of Proton is right to be concerned—but he’s also right that with the right habits and services, you can benefit from AI without surrendering your data.