AI chatbots such as ChatGPT, Claude and Copilot are growing increasingly popular, but cybersecurity company ESET warns users to think carefully before sharing personal, medical, financial or confidential information with these tools.
As artificial intelligence becomes part of everyday life, understanding what information to keep private has become an essential digital skill. ESET cautions that users may not always know how chatbot providers handle the information they share, particularly when they seek advice on mental health, medical conditions, legal matters or other personal issues.
Before choosing an AI platform or model, users should review the provider’s data policies and privacy settings to understand how it collects, stores and uses their information.
Although some platforms emphasise privacy and data sovereignty, users should remain cautious. Companies can change their policies, terms of service and priorities, potentially allowing information shared today to be used for AI training in the future.
Five Types of Information You Should Keep Private
ESET identifies five categories of information that users should avoid sharing with AI chatbots.
1. Passwords and account credentials
Users should never disclose passwords, login credentials, recovery codes or answers to security questions that could provide access to their accounts.
2. Sensitive personal information
Identity and passport details, home addresses and financial information can expose individuals to identity theft, fraud and other risks if they fall into the wrong hands.
3. Confidential workplace information
Internal reports, customer records, product plans, source code and other non-public business information should remain confidential unless an organisation has explicitly approved the AI tool for that purpose.
Uploading a confidential document for an AI-generated summary may seem like an efficient way to save time. However, if the tool lacks the necessary approval or appropriate agreements between the employer and provider, sharing the document could breach company policies, contractual obligations or data protection requirements.
4. Medical and highly sensitive personal data
Medical records, test results, treatment details and other health information require particular care because of their sensitive nature.
5. Information about other people
Users should avoid sharing colleagues’ details, customer records, private conversations, contact information and other personal data belonging to third parties without appropriate authorisation.
Before submitting any information, ESET recommends asking three questions: Is it publicly available? Am I authorised to share it? Would I feel comfortable if it became exposed in a data breach?
How to Use AI Chatbots More Safely
If users need to share information with an AI chatbot, ESET recommends taking steps to minimise potential risks.
These include removing names and other identifying details, using fictional or anonymised examples instead of real cases, and summarising sensitive documents themselves before asking AI for assistance.
At work, employees should use approved corporate AI tools with appropriate data protection safeguards whenever available.
Users should also check which classifications of information their chosen tools can process, whether conversation history and memory features are enabled, and whether external plugins, links or web-search functions could transfer information outside their organisation’s controlled environment.
Anonymisation and Deleting Chats May Not Be Enough
ESET warns that AI chatbots may infer personal information from a user’s prompt history, even when the user has anonymised the information entered.
Deleting a conversation also does not necessarily guarantee that all associated data or copies have disappeared. While deletion may reduce exposure, users should not assume that it removes every trace of the information.
Understanding each provider’s data retention policies and how it handles collected information is therefore essential, including whether it may use that information to train AI models.
What to Do If You Have Shared Sensitive Information
If users have already disclosed sensitive information, ESET recommends deleting the conversation where possible, removing any relevant stored memories and revoking access to connected applications where appropriate.
Anyone who has exposed passwords, recovery codes or other account credentials should change them promptly.
If the disclosure involves company information, customer records or data belonging to third parties, users should report the incident to their employer or organisational security team rather than attempting to manage it independently.
As AI chatbots become more widely used, ESET’s advice highlights the importance of treating personal and confidential information with care before submitting it to an AI tool.
(Source: AMNA)
The post Warnings Against Sharing Sensitive Information With AI Chatbots appeared first on Greek City Times.