Befriending an AI chatbot—could it mean compromising your data? How safe is it to share your personal information with an AI?

Have you ever shared your passwords, medical reports, or confidential office documents with an AI chatbot? If so, you should be cautious. The privacy of the information you share with an AI—while seeking answers to your questions—depends on your specific settings and the company’s policies. Data misuse, leaks, or unauthorized access can put your personal information at risk.

People now use AI tools like ChatGPT, Gemini, and Claude for a wide range of tasks—from studying and office work to drafting emails, coding, and even interpreting symptoms of an illness. Users often interact with these tools as if they were personal assistants or friends, sometimes even sharing private details. But how are your chats, uploaded files, and personal information actually used? Is your data truly secure?

Research published in October 2025 by Stanford University’s Institute for Human-Centered AI examined the privacy policies of six major American AI companies. It found that, by default, all six companies had settings in place that allowed user chats to be used for training and improving their AI models. However, the specific rules regarding data usage and the options available to users vary from one company to another.

What happens to your chats?
When you enter a query or upload a file to an AI chatbot, your data may be processed on the service provider’s servers. Depending on the company’s privacy policy and your personal settings, this data could be used to improve services, conduct security checks, or train AI models.

Cybersecurity expert Jiten Jain explains that, depending on the AI ​​provider’s privacy policy and settings, your data may be processed on their servers or by authorized service providers. Therefore, avoid uploading sensitive documents, passwords, financial records, or medical information that could reveal your identity unless absolutely necessary. AI systems may use conversations to improve their models, depending on the specific service, settings, and policies in place. There is always a risk of unauthorized access, data leaks, or the accidental exposure of personal information. Therefore, share only what is strictly necessary when interacting with a chatbot.

Not every piece of information provided to an AI is automatically incorporated into its model training. Furthermore, the fact that data is processed on a server does not mean it has become public. While privacy settings can help mitigate risks, they do not guarantee complete safety.

Why should data leaks serve as a warning?
SBI Data Leak: In October 2026, 3,897 records—out of a total of 375,000—containing customer names, phone numbers, email addresses, and physical addresses were discovered on the dark web. An investigation into how this leak occurred is currently underway.
Australian Government Portal: In June 2026, an OpenAI AI agent gained unauthorized access to a government Medicare statistics portal.
Samsung’s Confidential Code: In 2023, Samsung employees uploaded confidential source code to ChatGPT for debugging purposes. Following this incident, the company imposed restrictions on the use of generative AI.

These instances demonstrate that data risks arise not only from hacking but also from the act of sharing sensitive information with AI tools.

How to keep your data safe:
ChatGPT: Go to 'Data Controls' within the settings and disable the option "Improve the model for everyone." Additionally, "Temporary Chat" sessions do not appear in your chat history and are not used to improve the model; however, for security reasons, a copy may be retained for up to 30 days.
Claude: Under the 'Privacy' section in your settings, disable the option "Help improve our AI models." This prevents new conversations from being used for future model training.

Google Gemini: Check the 'Keep Activity' setting within your Gemini Apps Activity. Disabling this affects how your activity is saved and used; however, some data may still be retained for reasons related to security or service requirements.

Across all AI tools: Do not share passwords, bank details, Aadhaar numbers, medical reports, or confidential office documents. If using AI for work, ensure you use services approved by your company and adhere to their data policies.

While AI chatbots are useful, they should not be treated as a digital locker. Cybersecurity expert Jiten Jain advises that before sharing personal details, you should consider whether the task can be accomplished without providing that specific information. Sharing less data is the simplest way to minimize risk.

Disclaimer: This content has been sourced and edited from TV9. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.