Does a VPN protect your privacy when using ChatGPT or Claude? What network encryption can and can’t hide from AI

A Virtual Private Network (VPN) offers valuable protection by concealing your location from AI providers and shielding your internet activity from your internet service provider. However, this measure alone cannot prevent AI companies from reading, retaining, and learning from the content of your conversations. Even the strongest VPNs cannot hide the actual information you input, any files or pictures you upload, your account particulars, or any conversation history.

Any encryption measures only safeguard the connection between your device and the AI company's server, not the content transmitted through it. Moreover, most chatbots retain your exchanges indefinitely and may utilize them to enhance future models. To maintain privacy, it's crucial to manually disable conversation storage and opt-out of data training in platforms like ChatGPT and Claude.

Both OpenAI and Anthropic require user consent for data use, with default settings favoring data retention. Paid business plans generally offer better privacy guarantees. While a VPN is a helpful starting point for network security, it isn't sufficient to shield your privacy from AI companies. To ensure more private AI interactions, explore services like Proton AG's Lumo.

This service employs zero-access encryption, meaning only you can decrypt saved conversations, and it refrains from storing any data unless you explicitly grant permission. It operates on European servers governed by GDPR and Swiss privacy laws. For even greater privacy, consider running open-source models offline using tools such as Ollama or LM Studio.