AI & Cybersecurity Briefing · Updated October 2026

What Changed Since Launch

This strand tracks a moving field on purpose. Checked against the 14 August 2026 launch version, three things moved and three held steady — here's exactly what, with nothing smoothed over.

Corrected or sharpened

The research got re-verified against current sources; these are the facts that no longer read the same.

Corrected
EU AI Act high-risk deadline
Aug 2, 2026 — obligations "arriving in real time"→Deferred to Dec 2, 2027 (standalone high-risk) / Aug 2, 2028 (embedded in regulated products), per a Digital Omnibus that entered into force days before the original deadline. Transparency & general-purpose-AI provider rules still took effect on schedule in Aug 2026 — it's the heaviest compliance lift that moved, not all of it.
Sharpened
Five Eyes AI warning, dated
"mid-2026"→June 22, 2026 — the exact date is now confirmed, along with their own framing of the threat timeline: "not years, it is months."
Updated
Future of Life Institute signatories
133,000+ (early 2026)→134,000+ (as of March 2026) on the superintelligence-pause statement.
Held steady

Re-checked against current 2026 data and still standing as reported at launch.

✓4.8M unfilled cybersecurity roles — ISC2 again declined to publish a fresh 2025 gap figure.
✓87% of organizations reporting an AI-driven cyberattack in the past year.
✓41% of hiring managers cite AI/ML defense as the #1 skills need, still ahead of cloud security.
Tracks changes against the 14 August 2026 launch of "The Current Landscape" and "Capability Without Governance." Sources: EU Digital Omnibus on AI (July 2026) · Five Eyes joint advisory (June 2026) · Future of Life Institute signatory count (March 2026) · ISC2 2024/2025 Workforce Study · CybersecurityGuide.org (2026). Compiled October 2026.