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

2026-08-20T07:23:33Zd8f71b597d2c7698ea5b79f6b2208c604a80bcbad208fb5158b02f17080030df
AI agentsAI securityAndroid securityDNS parserGoogle Workspace GeminiHTTPSLLM securityLummaC2Merkle Tree CertificatesPQCRustTLSbrowser securitycellular baseband securitycertificate transparencydevice-bound session credentialsindirect prompt injectioninfostealer malwarememory safetypost-quantum cryptographyremote code executionsession cookie theft

What happened

Google security updates describe active and emerging threats including indirect prompt injection against AI agents and Workspace Gemini, session-cookie theft by infostealers, remote exploitation risks in cellular modem firmware, and the need to migrate Android and HTTPS infrastructure to post-quantum cryptography. The materials emphasize continuous defensive improvements, memory-safe Rust adoption, device-bound credentials, and quantum-resistant designs. No specific software vulnerability or CVE is identified.

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
google_security_blog
Record identifier
d8f71b597d2c7698ea5b79f6b2208c604a80bcbad208fb5158b02f17080030df
Enrichment time
2026-08-20T07:23:33Z
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.