Anthropic closes June with export restrictions lifted and a new mid-tier model shipping, while a novel prompt injection technique makes the case that AI-powered browsers may be structurally unsafe — r
Meta has confirmed that thousands of Instagram accounts were compromised through an attack vector that weaponized the platform's built-in AI chatbot — marking one of the first publicly confirmed large
Three security disclosures landed in the same day — a password manager breach with minimal explanation, a notification-delivered AI hijack, and wellness data stolen through an internal tool — while bo
Researchers at PromptArmor [documented a data exfiltration vulnerability](https://www.promptarmor.com/resources/gpt-for-google-sheets-data-exfiltration) in the popular "GPT for Google Sheets" add-on —
A developer has deliberately embedded a hidden prompt injection into **jqwik**, a widely-used Java property-based testing library, instructing AI coding agents to silently delete application output wh