ainewsblitz.com

Breaking

ChatGPT's Memory System Automatically Infers Your Personality, Habits, and Personal Data

  • Foundation Models
  • AI Agents
  • Security

OpenAI's ChatGPT memory feature—designed to personalize conversations by remembering user details—is automatically extracting sensitive personal information at scale, often without explicit user instruction or awareness. A comprehensive study at ETH Zurich analyzed 62,090 conversations and found that the system achieved approximately 61% accuracy in inferring personality traits from chat patterns alone, far exceeding random guessing. The findings underscore a broader architectural challenge: the memory system collects identifying information, medical records, financial details, and behavioral patterns that most users do not realize are being captured or permanently retained.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 26,424 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year