A virtual private network (VPN) can hide your real IP address from an AI platform and keep your internet service provider (ISP) from seeing that you’re using an AI service. But it does not hide the content you send to ChatGPT or Claude, according to TechRadar.
When a prompt reaches OpenAI or Anthropic, it must be decrypted so the service can process it. The companies can then see prompts, uploaded files or images, account details, conversation history and payment information. A VPN encrypts traffic between your device and its server; it does not control what happens after your data reaches the AI provider.
TechRadar says OpenAI may use ChatGPT content to improve its services unless users opt out in Settings > Data Controls. Chats may still be stored for safety and legal reasons. Anthropic asks consumer users whether their chats can be used for training, with the setting on by default. The article says data may be retained for up to five years when training is enabled, compared with 30 days when it is declined. It also notes that paid business tiers have stricter privacy agreements by default.
The article points to temporary chats in ChatGPT and Incognito mode in Claude as ways to keep conversations out of saved history. It also mentions Proton AG’s Lumo and offline open-source models run with Ollama or LM Studio as other approaches, while emphasizing that a VPN alone does not prevent AI providers from reading prompts.
Comments
0No comments yet. Be the first to comment.