Top 3 Most Notorious AI Data Privacy Fails

History is littered with high-profile AI-driven privacy lapses where training data or user content was exposed to the public.

The top 3

  1. The ChatGPT Chat History Exposure (2023): A software bug in the Redis library caused ChatGPT users to see the chat titles and metadata of other active users. OpenAI
  2. Samsung AI Data Leak via ChatGPT: Samsung employees inadvertently leaked sensitive source code and internal meeting notes by pasting them into ChatGPT for optimization. Samsung Electronics
  3. Microsoft AI Research Data Breach: A misconfigured Azure blob storage bucket exposed 38TB of private Microsoft research data, including training images for AI models. Microsoft

Sources

Open the full topic