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

Google security research highlights active and emerging threats across AI systems, browser session security, cellular modem firmware, and cryptographic infrastructure. Key topics include real-world indirect prompt injection against AI agents and Workspace integrations, device-bound session credentials to prevent stolen-cookie abuse, memory-safe Rust adoption in Pixel modem DNS parsing, and migration toward post-quantum cryptography for Android and HTTPS. These are defensive research and product security initiatives rather than disclosures of a specific exploitable vulnerability.