Are Your AI Chatbot Conversations Truly Private?

AI-generated, human-reviewed.

Most people don’t realize that conversations with AI chatbots like ChatGPT and Gemini are not as private as they think. On Hands-On Tech, Mikah Sargent uncovered how these services store and use your chats—and what you can do to protect your data.

Quick Summary

Whenever you use an AI chatbot, your prompts, uploads, and even voice recordings are saved to company servers and often tied to your account. These records may be accessed by human reviewers and are, by default, used to train and improve the AI. While there are settings to limit training, data storage usually continues—and deleted chats may not be fully gone.

Privacy controls do exist on both ChatGPT and Gemini, but each handles chat history and data retention differently. Knowing how to adjust these settings is key for anyone entering sensitive information.

Where Does Your Chatbot Data Go?

AI chatbots store every interaction—including typed messages, uploads, voice, and images—on their servers. This is why chat history “follows” you from device to device. By default:

• ChatGPT (OpenAI) saves chats to your account
• Gemini (Google) also stores chat and device data connected to your account

For most personal accounts, these conversations help train the AI, which means your private data can help improve future chatbot responses.

Who Can Access Your Conversations?

• Human reviewers: Both Google and OpenAI explicitly state that authorized personnel can review chats for support, abuse investigation, legal compliance, or model improvement.
• Third parties: In legal matters, data may be shared with courts or authorities.
• Advertisers: While conversations themselves aren’t usually sold to advertisers, certain plans allow ad personalization based on your chat history.

Critically, there’s no doctor-patient or attorney-client privilege protecting what you share with an AI. Even health data is only covered by company policies, not federal privacy laws like HIPAA.

How to Manage and Protect Your Privacy

On Hands-On Tech, Mikah Sargent demonstrated that both ChatGPT and Gemini offer privacy controls, but they come with trade-offs.

ChatGPT (OpenAI):

• To turn off training:
  • Go to Settings > Data Controls > toggle off “Improve the Model for Everyone”
  • Chats remain in your history but are not used for training.
  • Temporary chat feature: Keeps chats out of history and memory, but OpenAI retains a copy for up to 30 days.
• To manage memories:
  • Settings > Personalization > Memory, where you can view and edit what ChatGPT has saved about you.
• Ad personalization: Free and Go plans can use chat data for personalization unless you opt out.

Gemini (Google):

• To stop model training:
  • In the Gemini app, go to your profile, Gemini Apps Activity, and turn off or delete activity.
  • Note: Disabling this also wipes chat history, making each conversation temporary.
• Auto-delete: Default is 18 months; can be set as low as 3 months.
• Even with activity off, chats may be stored for up to 72 hours for processing.
• Temporary chats: Not used for training or human review, with a 72-hour retention window.

Key Takeaways

• AI chatbot data is stored and can be reviewed—sometimes for years.
• Turning off training does not turn off storage, only model improvement.
• Chats are not confidential: Sensitive info like health records, names, or addresses should be omitted or redacted.
• OpenAI’s ChatGPT Health offers extra protections and separate memory for medical questions—but is still not covered by federal health privacy laws.
• Be cautious: Deleting a chat does not always delete the associated memory or guarantee removal from backups/legal holds.

What This Means for You

When you use AI chatbots for anything sensitive—health, legal, finances—always treat the conversation as something the company (and potentially others) can access. Use dedicated features like ChatGPT Health for medical matters and temporary chat modes for privacy. Above all, never assume these tools provide the same privacy as a confidential conversation with a professional.

The Bottom Line

AI chatbots are powerful and helpful, but your data is almost always stored and potentially reviewed unless you take specific steps. For the best privacy, turn off training, use temporary chats, and avoid sharing personally identifiable information whenever possible.

Want more practical tech advice? Subscribe to Hands-On Tech on TWiT.tv:
https://twit.tv/shows/hands-on-tech/episodes/285