A security researcher has uncovered evidence that user data, including prompts, is being leaked from multiple AI providers. Attackers reportedly gained access to inference marketplaces, siphoning prompts from at least three AI providers. Additionally, some AI providers and resellers were found to disregard user privacy settings like Zero Data Retention, exposing sensitive information. AI
IMPACT Exposes user data and prompts, undermining trust in AI privacy settings and potentially leading to misuse of sensitive information.
RANK_REASON Security vulnerability impacting multiple AI products and services.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →