Before You Paste Anything Into ChatGPT, Check This List

When considering whether to input sensitive information into an AI chatbot like ChatGPT, it is crucial to carefully evaluate the nature of the data and the potential risks involved. Certain types of information should never be shared with AI platforms, regardless of the convenience they offer. This includes passwords, medical records, source code, company data, credentials, financial details, personal identifiers, and proprietary information.

Even seemingly harmless details like a passport number or driver's license number can create vulnerabilities if exposed. Additionally, internal financial results, customer records, business processes, and unreleased information should not be submitted to unapproved AI services. Developers, in particular, must exercise caution when sharing source code, as it may contain sensitive elements such as API keys, database passwords, and proprietary algorithms.

Medical information, while potentially useful, should be approached with caution due to the possibility of AI hallucinations leading to incorrect advice. Before pasting any data into an AI chatbot, ask yourself if the information is truly necessary for the chatbot to assist you. If in doubt, it is safer to refrain from uploading the information altogether.

Remember that AI chatbots process and handle the submitted data, unlike a private diary stored solely on your device. While there have been no confirmed cases of major AI platforms being breached solely due to chatbot conversations, the potential for data theft and legal access issues should not be underestimated. AI conversations generally lack confidentiality protections, and government agencies may require providers to disclose user data.

Ultimately, the safest assumption is that any information submitted to an AI chatbot is being processed by a service with potentially differing privacy, retention, access, and security rules from your own expectations.