# NSA guidance addresses generative AI deepfakes and content authenticity

Published: 2025-01-29 · Severity: routine · Sectors: government-national, financial-services, technology, healthcare
Canonical: https://vorant.io/reports/b75f0645-49fb-5525-9396-060673df2999/nsa-guidance-addresses-generative-ai-deepfakes-and-content-authenticity

> NSA and NCSC endorse Content Credentials technology to counter deepfakes, impersonation, and synthetic media in AI-driven attacks.

The NCSC and NSA have published guidance addressing the erosion of trust caused by generative AI and deepfake technologies. Attackers increasingly use AI to create convincing synthetic content—text, images, video, and voice—for impersonation, credential theft, and spear-phishing attacks. The FBI warned in December 2024 of criminals using AI-generated documents and images in private communications to impersonate legitimate individuals. Content Credentials is an emerging technology designed to establish and verify data lineage, source attribution, and editing history to authenticate content and detect synthetic material.

The guidance highlights several technical challenges: detection tools for synthetic content are often unreliable, AI systems suffer from 'model collapse' when trained on previously generated data (causing compounding errors), and attackers exploit these weaknesses to craft more convincing fraudulent communications. The NCSC and international partners endorse Content Credentials and watermarking standards as layered defences, though no single technical measure will fully solve the problem. Organisations should begin preparing now to implement content provenance techniques, particularly for high-value use cases like sensitive communications, media verification, and authentication systems.

Source reporting: https://www.ncsc.gov.uk/blog-post/preserving-integrity-in-age-generative-ai

---

This is the free public brief from Vorant Threat Intelligence. When citing, attribute "Vorant" and link https://vorant.io/reports/b75f0645-49fb-5525-9396-060673df2999/nsa-guidance-addresses-generative-ai-deepfakes-and-content-authenticity.
In the app the same report carries its extracted indicators, its detections (Splunk SPL already written, plus Elastic and Microsoft KQL where they apply), live profiles of the actors and CVEs it names, and the vendor research on the same campaign. Slack alerts fire on the vendors, sectors and countries a reader follows. A new account starts with three days of all of it, no card: https://vorant.io/signup