0 / 0
Personal information in prompt risk for AI

Personal information in prompt risk for AI

Privacy Icon representing privacy risks.
Risks associated with input
Inference
Privacy
New to generative AI

Description

Personal information or sensitive personal information that is included as a part of a prompt that is sent to the model.

Why is personal information in prompt a concern for foundation models?

If personal information or sensitive personal information is included in the prompt, it might be unintentionally disclosed in the models’ output. In addition to accidental disclosure, prompt data might be stored or later used for other purposes like model evaluation and retraining, and might appear in their output if not properly removed. 

Background image for risks associated with input
Example

Disclose personal health information in ChatGPT prompts

According to the source article, some people on social media shared about using ChatGPT as their makeshift therapists. Articles contend that users might include personal health information in their prompts during the interaction, which might raise privacy concerns. The information might be shared with the company that own the technology and might be used for training or tuning or even shared with unspecified third parties.

Parent topic: AI risk atlas

We provide examples covered by the press to help explain many of the foundation models' risks. Many of these events covered by the press are either still evolving or have been resolved, and referencing them can help the reader understand the potential risks and work towards mitigations. Highlighting these examples are for illustrative purposes only.

Generative AI search and answer
These answers are generated by a large language model in watsonx.ai based on content from the product documentation. Learn more