Grok Image Edit Reportedly Leaks Original Images and Prompts
mark_k · x · 2026-08-03
A user discovered that X's Grok Imagine not only leaks the prompts used for generated content, but also exposes the original source images during image edits.
By clicking the label under the image, anyone can easily access the underlying photo. This could inadvertently reveal sensitive information, such as the real identities of the people in the picture, posing a major privacy risk.
Related event: X's Grok Imagine Exposes User Prompts and Original Images(3 posts)→
More from Safety
- OpenAI Disrupts Cambodia-Based Criminal Scam Operation Using ChatGPT — OpenAI News · 2026-08-04
- AI Agents Are Breaking Double-Blind Peer Review in Academia — deliprao · 2026-08-03
- New Paper Analyzes Quasi-AI Regulation Through Exclusive Staged Access for Cybersecurity — joshgans · 2026-08-03
- Facebook Caught Uploading Unposted Photos for AI Analysis — ahuja_priyank · 2026-08-03
- AI Therapy Chatbots Spark Patient Safety Fears, Prompting State Regulations — bloomberglaw · 2026-08-03
- Google DeepMind proposes framework for intelligent AI delegation to secure agentic web — rvp · 2026-08-03