AI threats in the wild: The current state of prompt injections on the web

Google Security Blog posts cover emerging security topics including real-world indirect prompt injection against AI agents and Workspace Gemini, memory-safety hardening of Pixel modem firmware using Rust, device-bound session credentials to mitigate browser cookie theft, vulnerability reward program results, and post-quantum cryptography adoption across Android and HTTPS. The most immediate threat themes are indirect prompt injection and infostealer-enabled session hijacking; the other posts describe defensive engineering and cryptographic migration initiatives.