Updated
Updated · ZDNet · Aug 20
Incogni Ranks 13 AI Platforms on Privacy Risk, Putting Copilot Last and Vibe First
Updated
Updated · ZDNet · Aug 20

Incogni Ranks 13 AI Platforms on Privacy Risk, Putting Copilot Last and Vibe First

2 articles · Updated · ZDNet · Aug 20

Summary

  • Incogni’s 2026 study scored 13 generative AI platforms on consumer privacy risk and found bigger platforms generally posed higher risks, with ChatGPT the main exception.
  • Vibe ranked safest, followed by ChatGPT and Pi, while Copilot ranked worst after Meta and Kimi; Gemini, Claude and Grok also landed in the higher-risk half.
  • The ranking weighed 3 areas: how user data is collected and shared, whether prompts feed model training and can be opted out of, and how clear privacy policies are.
  • Copilot was flagged for sharing data with third-party advertisers and buying data from brokers, while Meta was labeled the most data-hungry and Gemini criticized for relying on broad Google privacy terms.
  • No platform lets users remove data once it has trained a model, Incogni said, reinforcing advice to avoid entering passwords, financial records, medical data or confidential work information.

Insights

Why are the world's most popular AI platforms quietly harvesting your biometric data and sharing it with unseen advertisers?
Are enterprise AI tools secretly exposing your company's most confidential internal files to anyone who knows exactly how to ask?
Could your casual conversation with an emotionally intelligent AI permanently trap your deepest secrets in an unerasable digital vault?