AI threats in the wild: The current state of prompt injections on the web
2026-06-22T13:23:40Z•1e8c676f95aaa6e6a5d689dfa6dce03b1ba2b2ce29b12ab0a20d4bf98e7ef10a
AI agent securityAndroidAndroid 17 betaChrome 146DBSCDNS parserGeminiGoogle WorkspaceIPILLM securityLumamaC2Merkle Tree Certificates (MTC)PQCRustcookie theftdevice bound session credentialsindirect prompt injectioninfostealermemory safetymodem securitypixel basebandpost-quantum cryptographyprompt injectionquantum-safe TLSsession theft
What happened
Collection of Google Security Blog posts (Apr–Mar 2026) covering multiple security topics: a broad web sweep found real-world activity and growing risk from Indirect Prompt Injection (IPI) targeting AI agents and LLM-based workflows, and Google describes continuous, multi-layered mitigations (including Workspace-specific defenses). Chrome is rolling out Device Bound Session Credentials (DBSC) to proactively prevent cookie/session theft from infostealer malware (Windows available in Chrome 146; macOS coming). Pixel team reports integrating a memory-safe Rust DNS parser into the modem/baseband (
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- google_security_blog
- Record identifier
- 1e8c676f95aaa6e6a5d689dfa6dce03b1ba2b2ce29b12ab0a20d4bf98e7ef10a
- Enrichment time
- 2026-06-22T13:23:40Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.